Apple Platform Security Guide

Apple
Platform Security
February 2021
Contents
Introduction to Apple platform security 5
A commitment to security 6
Hardware security and biometrics 7
Hardware security overview 7

Apple SoC security 8
Secure Enclave 9
Touch ID and Face ID 17
Hardware microphone disconnect 23
Express Cards with power reserve 24
System security 25
System security overview 25

Secure boot 25
Secure software updates 48
Operating system integrity 50
Additional macOS system security capabilities 53
System security for watchOS 65
Random number generation 68
Apple Security Research Device 68
Encryption and Data Protection 70
Encryption and Data Protection overview 70

Passcodes and passwords 70
Data Protection 73
FileVault 87
How Apple protects users’ personal data 90
Digital signing and encryption 92
Apple Platform Security 2

App security 95
App security overview 95

App security in iOS and iPadOS 96
App security in macOS 101
Secure features in the Notes app 105
Secure features in the Shortcuts app 106
Services security 107
Services security overview 107

Apple ID and Managed Apple ID 107
iCloud 109
Passcode and password management 112
Apple Pay 121
iMessage 134
Secure Business Chat using the Messages app 137
FaceTime security 138
Find My 139
Continuity 143
Car keys security in iOS 146
Network security 149
Network security overview 149

TLS security 149
IPv6 security 151
Virtual private network (VPN) security 152
Wi-Fi security 153
Bluetooth security 157
Ultra Wideband security in iOS 158
Single sign-on security 159
AirDrop security 160
Wi-Fi password sharing security on iPhone and iPad 161
Firewall security in macOS 162

Developer kit security 163
Developer kit security overview 163

HomeKit 163
CloudKit security 170
SiriKit security for iOS, iPadOS, and watchOS 170
DriverKit security for macOS 10.15 171
ReplayKit security in iOS and iPadOS 171
ARKit security in iOS and iPadOS 173
Secure device management 174
Secure device management overview 174

Pairing model security for iPhone and iPad 174
Mobile device management 175
Apple Configurator 2 security 184
Screen Time security 185
Glossary 187
Document revision history 192

Introduction to Apple platform security
Apple designs security into the core of its platforms. Building on the experience of
creating the world’s most advanced mobile operating system, Apple has created security
architectures that address the unique requirements of mobile, watch, desktop, and home.
Every Apple device combines hardware, software, and services designed to work together
for maximum security and a transparent user experience in service of the ultimate goal
of keeping personal information safe. For example, Apple-designed silicon and security
hardware powers critical security features. And software protections work to keep the
operating system and third-party apps protected. Finally, services provide a mechanism
for secure and timely software updates, power a protected app ecosystem, and facilitate
secure communications and payments. As a result, Apple devices protect not only the
device and its data but the entire ecosystem, including everything users do locally, on
networks, and with key internet services.
Just as we design our products to be simple, intuitive, and capable, we design them to
be secure. Key security features, such as hardware-based device encryption, can’t be
disabled by mistake. Other features, such as Touch ID and Face ID, enhance the user
experience by making it simpler and more intuitive to secure the device. And because many
of these features are enabled by default, users or IT departments don’t need to perform
extensive configurations.
This documentation provides details about how security technology and features are
implemented within Apple platforms. It also helps organizations combine Apple platform
security technology and features with their own policies and procedures to meet their
specific security needs.
The content is organized into the following topic areas:
• Hardware security and biometrics: The silicon and hardware that forms the foundation
for security on Apple devices, including the Secure Enclave, a dedicated AES
cryptographic engine, Touch ID, and Face ID
• System security: The integrated hardware and software functions that provide for the
safe boot, update, and ongoing operation of Apple operating systems
• Encryption and Data Protection: The architecture and design that protects user data
if the device is lost or stolen or if an unauthorized person or process attempts to use or
modify it
• App security: The software and services that provide a safe app ecosystem and enable
apps to run securely and without compromising platform integrity
• Services security: Apple’s services for identification, password management, payments,

communications, and finding lost devices

• Network security: Industry-standard networking protocols that provide secure
authentication and encryption of data in transmission
• Developer kit security: Framework “kits” for secure and private management of home
and health, as well as extension of Apple device and service capabilities to third-party
apps
• Secure device management: Methods that allow management of Apple devices,

prevent unauthorized use, and enable remote wipe if a device is lost or stolen
A commitment to security
Apple is committed to helping protect customers with leading privacy and security
technologies—designed to safeguard personal information—and comprehensive
methods, to help protect corporate data in an enterprise environment. Apple
rewards researchers for the work they do to uncover vulnerabilities by offering the
Apple Security Bounty. Details of the program and bounty categories are available at
https://developer.apple.com/security-bounty/.
We maintain a dedicated security team to support all Apple products. The team provides
security auditing and testing for products, both under development and released. The
Apple team also provides security tools and training, and actively monitors for threats and
reports of new security issues. Apple is a member of the Forum of Incident Response and
Security Teams (FIRST).
Apple continues to push the boundaries of what’s possible in security and privacy. This
year Apple devices with Apple SoC’s across the product lineup from Apple Watch to iPhone
and iPad, and now Mac, utilize custom silicon to power not only efficient computation,
but also security. Apple silicon forms the foundation for secure boot, Touch ID and
Face ID, and Data Protection, as well as system integrity features never before featured
on the Mac including Kernel Integrity Protection, Pointer Authentication Codes, and Fast
Permission Restrictions. These integrity features help prevent common attack techniques
that target memory, manipulate instructions, and use javascript on the web. They combine
to help make sure that even if attacker code somehow executes, the damage it can do is
dramatically reduced.
To make the most of the extensive security features built into our platforms, organizations
are encouraged to review their IT and security policies to ensure that they are taking full
advantage of the layers of security technology offered by these platforms.
To learn more about reporting issues to Apple and subscribing to security notifications, see
Report a security or privacy vulnerability.
Apple believes privacy is a fundamental human right and has numerous built-in
controls and options that allow users to decide how and when apps use their
information, as well as what information is being used. To learn more about Apple’s
approach to privacy, privacy controls on Apple devices, and the Apple privacy policy,
see https://www.apple.com/privacy.
Note: Unless otherwise noted, this documentation covers the following operating system
versions: iOS 14.3, iPadOS 14.3, macOS 11.1, tvOS 14.3, and watchOS 7.2.

Hardware security and biometrics
Hardware security overview

For software to be secure, it must rest on hardware that has security built in. That’s why
Apple devices—running iOS, iPadOS, macOS, watchOS, or tvOS—have security capabilities
designed into silicon. These capabilities include a CPU that powers system security
features, as well as additional silicon that’s dedicated to security functions. Security-
focused hardware follows the principle of supporting limited and discretely defined
functions in order to minimize attack surface. Such components include a boot ROM,
which forms a hardware root of trust for secure boot, dedicated AES engines for efficient
and secure encryption and decryption, and a Secure Enclave. The Secure Enclave is a
system on chip (SoC) that is included on all recent iPhone, iPad, Apple Watch, Apple TV
and HomePod devices, and on a Mac with Apple silicon as well as those with the Apple T2
Security Chip. The Secure Enclave itself follows the same principle of design as the SoC
does, containing its own discrete boot ROM and AES engine. The Secure Enclave also
provides the foundation for the secure generation and storage of the keys necessary for
encrypting data at rest, and it protects and evaluates the biometric data for Touch ID and
Face ID.
Storage encryption must be fast and efficient. At the same time, it can’t expose the data
(or keying material) it uses to establish cryptographic keying relationships. The AES
hardware engine solves this problem by performing fast in-line encryption and decryption
as files are written or read. A special channel from the Secure Enclave provides necessary
keying material to the AES engine without exposing this information to the Application
Processor (or CPU) or overall operating system. This ensures that the Apple Data
Protection and FileVault technologies protect users’ files without exposing long-lived
encryption keys.
Apple has designed secure boot to protect the lowest levels of software against tampering
and to allow only trusted operating system software from Apple to load at startup. Secure
boot begins in immutable code called the Boot ROM, which is laid down during Apple SoC
fabrication and is known as the hardware root of trust. On Mac computers with a T2 chip,
trust for macOS secure boot begins with the T2. (Both the T2 chip and the Secure Enclave
also execute their own secure boot processes using their own separate boot ROM—this is
an exact analogue to how the A-series and M1 chips boot securely.)
The Secure Enclave also processes fingerprint and face data from Touch ID and Face ID
sensors in Apple devices. This provides secure authentication while keeping user biometric
data private and secure. It also enables users to benefit from the security of longer and
more complex passcodes and passwords with, in many situations, the convenience of swift
authentication for access or purchases.

Apple SoC security
Apple-designed silicon forms a common architecture across all Apple products and
now powers Mac as well as iPhone, iPad, Apple TV, and Apple Watch. For over a decade,
Apple’s world-class silicon design team has been building and refining Apple systems on
chip (SoCs). The result is a scalable architecture designed for all devices that leads the
industry in security capabilities. This common foundation for security features is only
possible from a company that designs its own silicon to work with its software.
Apple silicon has been designed and fabricated to specifically enable the system security
features detailed below.
Feature A10 A11, S3 A12, S4 A13, S5 A14, S6 M1
Kernel
Integrity
Protection
Fast
Permission
Restrictions
System
Coprocessor
Integrity
Protection
Pointer
Authentication
Codes
Page See Note

Protection below.
Layer
Note: Page Protection Layer (PPL) requires that the platform execute only signed and
trusted code; this is a security model that isn’t applicable on macOS.
Apple-designed silicon also specifically enables the Data Protection capabilities detailed
below.
Feature A10 A11, S3 A12, S4 A13, S5 A14, M1, S6
Sealed Key
Protection (SKP)
recoveryO S - All
Data Protection
Classes
protected
Alternate
boots of DFU,
Diagnostics, and
Update - Class
A, B, and C data
protected

Secure Enclave
Overview
The Secure Enclave is a dedicated secure subsystem integrated into Apple systems on chip
(SoCs). The Secure Enclave is isolated from the main processor to provide an extra layer
of security and is designed to keep sensitive user data secure even when the Application
Processor kernel becomes compromised. It follows the same design principles as the SoC
does—a boot ROM to establish a hardware root of trust, an AES engine for efficient and
secure cryptographic operations, and protected memory. Although the Secure Enclave
doesn’t include storage, it has a mechanism to store information securely on attached
storage separate from the NAND flash storage that’s used by the Application Processor
and operating system.
The Secure Enclave components.

The Secure Enclave is a hardware feature of most versions of iPhone, iPad, Mac, Apple TV,
Apple Watch, and HomePod—namely:
• iPhone 5s or later
• iPad Air or later
• MacBook Pro computers with Touch Bar (2016 and 2017) that contain the Apple T1 Chip
• Intel-based Mac computers that contain the Apple T2 Security Chip
• Mac computers with Apple silicon
• Apple TV HD or later
• Apple Watch Series 1 or later
• HomePod and HomePod mini
Secure Enclave Processor

The Secure Enclave Processor provides the main computing power for the Secure Enclave.
To provide the strongest isolation, the Secure Enclave Processor is dedicated solely for
Secure Enclave use. This helps prevent side-channel attacks that depend on malicious
software sharing the same execution core as the target software under attack.
The Secure Enclave Processor runs an Apple-customized version of the L4 microkernel. It’s
designed to operate efficiently at a lower clock speed that helps to protect it against clock
and power attacks. The Secure Enclave Processor, starting with the A11 and S4, includes
a memory-protected engine and encrypted memory with anti-replay capabilities, secure
boot, a dedicated random number generator, and its own AES engine.
Memory Protection Engine

The Secure Enclave operates from a dedicated region of the the device’s DRAM memory.
Multiple layers of protection isolate the Secure Enclave protected memory from the
Application Processor.
When the device starts up, the Secure Enclave Boot ROM generates a random ephemeral
memory protection key for the Memory Protection Engine. Whenever the Secure Enclave
writes to its dedicated memory region, the Memory Protection Engine encrypts the block
of memory using AES in Mac XEX (xor-encrypt-xor) mode, and calculates a Cipher-
based Message Authentication Code (CMAC) authentication tag for the memory. The
Memory Protection Engine stores the authentication tag alongside the encrypted memory.
When the Secure Enclave reads the memory, the Memory Protection Engine verifies the
authentication tag. If the authentication tag matches, the Memory Protection Engine
decrypts the block of memory. If the tag doesn’t match, the Memory Protection Engine
signals an error to the Secure Enclave. After a memory authentication error, the Secure
Enclave stops accepting requests until the system is rebooted.

Starting with the Apple A11 and S4 SoCs, the Memory Protection Engine adds replay
protection for Secure Enclave memory. To prevent replay of security-critical data, the
Memory Protection Engine stores a nonce for the block of memory alongside the
authentication tag. The nonce is used as an additional tweak for the CMAC authentication
tag. The nonces for all memory blocks are protected using an integrity tree rooted in
dedicated SRAM within the Secure Enclave. For writes, the Memory Protection Engine
updates the nonce and each level of the integrity tree up to the SRAM. For reads, the
Memory Protection Engine verifies the nonce and each level of the integrity tree up to the
SRAM. Nonce mismatches are handled similarly to authentication tag mismatches.
On Apple A14, M1, and later SoCS, the Memory Protection Engine supports two ephemeral
memory protection keys. The first is used for data private to the Secure Enclave, and the
second is used for data shared with the Secure Neural Engine.
The Memory Protection Engine operates inline and transparently to the Secure
Enclave. The Secure Enclave reads and writes memory as if it were regular unencrypted
DRAM, whereas an observer outside the Secure Enclave sees only the encrypted and
authenticated version of the memory. The result is strong memory protection without
performance or software complexity tradeoffs.
Secure Enclave Boot ROM

The Secure Enclave includes a dedicated Secure Enclave Boot ROM. Like the Application
Processor Boot ROM, the Secure Enclave Boot ROM is immutable code that establishes the
hardware root of trust for the Secure Enclave.
On system startup, iBoot assigns a dedicated region of memory to the Secure Enclave.
Before using the memory, the Secure Enclave Boot ROM initializes the Memory Protection
Engine to provide cryptographic protection of the Secure Enclave protected memory.
The Application Processor then sends the sepOS image to the Secure Enclave Boot ROM.
After copying the sepOS image into the Secure Enclave protected memory, the Secure
Enclave Boot ROM checks the cryptographic hash and signature of the image to verify that
the sepOS is authorized to run on the device. If the sepOS image is properly signed to run
on the device, the Secure Enclave Boot ROM transfers control to sepOS. If the signature
isn’t valid, the Secure Enclave Boot ROM prevents any further use of the Secure Enclave
until the next chip reset.
On Apple A10 and later SoCs, the Secure Enclave Boot ROM locks a hash of the sepOS
into a register dedicated to this purpose. The Public Key Accelerator uses this hash for
operating-system-bound (OS-bound) keys.

Secure Enclave Boot Monitor
On Apple A13 and later SoCs, the Secure Enclave includes a Boot Monitor to ensure
stronger integrity on the hash of the booted sepOS.
At system startup, the Secure Enclave Processor’s System Coprocessor Integrity

Protection (SCIP) configuration prevents the Secure Enclave Processor from executing
any code other than the Secure Enclave Boot ROM. The Boot Monitor prevents the Secure
Enclave from modifying the SCIP configuration directly. To make the loaded sepOS
executable, the Secure Enclave Boot ROM sends the Boot Monitor a request with the
address and size of the loaded sepOS. On receipt of the request, the Boot Monitor resets
the Secure Enclave Processor, hashes the loaded sepOS, updates the SCIP settings to
allow execution of the loaded sepOS, and starts execution within the newly loaded code.
As the system continues booting, this same process is used whenever new code is made
executable. Each time, the Boot Monitor updates a running hash of the boot process. The
Boot Monitor also includes critical security parameters in the running hash.
When boot completes, the Boot Monitor finalizes the running hash and sends it to the
Public Key Accelerator to use for OS-bound keys. This process is designed so that
operating system key binding can’t be bypassed even with a vulnerability in the Secure
Enclave Boot ROM.
True Random Number Generator

The True Random Number Generator (TRNG) is used to generate secure random data. The
Secure Enclave uses the TRNG whenever it generates a random cryptographic key, random
key seed, or other entropy. The TRNG is based on multiple ring oscillators post processed
with CTR_DRBG (an algorithm based on block ciphers in Counter Mode).
Root Cryptographic Keys

The Secure Enclave includes a unique ID (UID) root cryptographic key. The UID is unique to
each individual device and isn’t related to any other identifier on the device.
A randomly generated UID is fused into the SoC at manufacturing time. Starting with A9
SoCs, the UID is generated by the Secure Enclave TRNG during manufacturing and written
to the fuses using a software process that runs entirely in the Secure Enclave. This process
protects the UID from being visible outside the device during manufacturing and therefore
isn’t available for access or storage by Apple or any of its suppliers.
sepOS uses the UID to protect device-specific secrets. The UID allows data to be
cryptographically tied to a particular device. For example, the key hierarchy protecting
the file system includes the UID, so if the internal SSD storage is physically moved from
one device to another, the files are inaccessible. Other protected device-specific secrets
include Touch ID or Face ID data. On a Mac, only fully internal storage linked to the AES
engine receives this level of encryption. For example, neither external storage devices
connected over USB nor PCIe-based storage added to the 2019 Mac Pro are encrypted in
this fashion.
The Secure Enclave also has a device group ID (GID), which is common to all devices that
use a given SoC (for example, all devices using the Apple A14 SoC share the same GID).
The UID and GID aren’t available through Joint Test Action Group (JTAG) or other
debugging interfaces.

Secure Enclave AES Engine
The Secure Enclave AES Engine is a hardware block used to perform symmetric
cryptography based on the AES cipher. The AES Engine is designed to resist leaking
information by using timing and Static Power Analysis (SPA). Starting with the A9 SoC, the
AES Engine also includes Dynamic Power Analysis (DPA) countermeasures.
The AES Engine supports hardware and software keys. Hardware keys are derived from
the Secure Enclave UID or GID. These keys stay within the AES Engine and aren’t made
visible even to sepOS software. Although software can request encryption and decryption
operations with hardware keys, it can’t extract the keys.
On Apple A10 and newer SoCs, the AES Engine includes lockable seed bits that diversify
keys derived from the UID or GID. This allows data access to be conditioned on the
device’s mode of operation. For example, lockable seed bits are used to deny access to
password-protected data when booting from Device Firmware Update (DFU) mode. For
more information, see Passcodes and passwords.
AES Engine
Every Apple device with a Secure Enclave also has a dedicated AES256 crypto engine
(the “AES Engine”) built into the direct memory access (DMA) path between the NAND
(nonvolatile) flash storage and main system memory, making file encryption highly efficient.
On A9 or later A-series processors, the flash storage subsystem is on an isolated bus
that’s granted access only to memory containing user data through the DMA crypto engine.
At boot time, sepOS generates an ephemeral wrapping key using the TRNG. The Secure
Enclave transmits this key to the AES Engine using dedicated wires to prevent it from
being accessed by any software outside the Secure Enclave. sepOS can then use the
ephemeral wrapping key to wrap file keys for use by the Application Processor file-system
driver. When the file-system driver reads or writes a file, it sends the wrapped key to the
AES Engine, which unwraps the key. The AES Engine never exposes the unwrapped key to
software.
Note: The AES Engine is a separate component from both the Secure Enclave and the
Secure Enclave AES Engine, but its operation is closely tied to the Secure Enclave, as
shown below.
The AES Engine supports line-speed encryption on the DMA path for efficient encryption and decryption of data is
it is written and read to storage.

Public Key Accelerator
The Public Key Accelerator (PKA) is a hardware block used to perform asymmetric
cryptography operations. The PKA supports RSA and ECC (Elliptic Curve Cryptography)
signing and encryption algorithms. The PKA is designed to resist leaking information using
timing and side-channel attacks such as SPA and DPA.
The PKA supports software and hardware keys. Hardware keys are derived from the Secure
Enclave UID or GID. These keys stay within the PKA and aren’t made visible even to sepOS
software.
Starting with A13 SoCs, the PKA’s encryption implementations have been proved to be
mathematically correct using formal verification techniques.
On Apple A10 and later SoCs, the PKA supports OS-bound keys, also referred to as Sealed
Key Protection (SKP). These keys are generated using a combination of the device’s UID
and the hash of the sepOS running on the device. The hash is provided by the Secure
Enclave Boot ROM, or by the Secure Enclave Boot Monitor on Apple A13 and later SoCs.
These keys are also used to verify the sepOS version when making requests to certain
Apple services and are also used to improve the security of passcode-protected data by
helping to prevent access to keying material if critical changes are made to the system
without user authorization.
Secure nonvolatile storage

The Secure Enclave is equipped with a dedicated secure nonvolatile storage device. The
secure nonvolatile storage is connected to the Secure Enclave using a dedicated I2C bus,
so that it can only be accessed by the Secure Enclave. All user data encryption keys are
rooted in entropy stored in the Secure Enclave nonvolatile storage.
In devices with A12, S4, and later SoCs, the Secure Enclave is paired with a Secure
Storage Component for entropy storage. The Secure Storage Component is itself
designed with immutable ROM code, a hardware random number generator, a per-device
unique cryptographic key, cryptography engines, and physical tamper detection. The
Secure Enclave and Secure Storage Component communicate using an encrypted and
authenticated protocol that provides exclusive access to the entropy.
Devices first released in Fall 2020 or later are equipped with a 2nd-generation Secure
Storage Component. The 2nd-generation Secure Storage Component adds counter
lockboxes. Each counter lockbox stores a 128-bit salt, a 128-bit passcode verifier, an 8-bit
counter, and an 8-bit maximum attempt value. Access to the counter lockboxes is through
an encrypted and authenticated protocol.
Counter lockboxes hold the entropy needed to unlock passcode-protected user data. To
access the user data, the paired Secure Enclave must derive the correct passcode entropy
value from the userʼs passcode and the Secure Enclaveʼs UID. The user’s passcode can’t
be learned using unlock attempts sent from a source other than the paired Secure Enclave.
If the passcode attempt limit is exceeded (for example, 10 attempts on iPhone), the
passcode-protected data is erased completely by the Secure Storage Component.

To create a counter lockbox, the Secure Enclave sends the Secure Storage Component the
passcode entropy value and the maximum attempt value. The Secure Storage Component
generates the salt value using its random number generator. It then derives a passcode
verifier value and a lockbox entropy value from the provided passcode entropy, the Secure
Storage Component’s unique cryptographic key, and the salt value. The Secure Storage
Component initializes the counter lockbox with a count of 0, the provided maximum
attempt value, the derived passcode verifier value, and the salt value. The Secure Storage
Component then returns the generated lockbox entropy value to the Secure Enclave.
To retrieve the lockbox entropy value from a counter lockbox later, the Secure Enclave
sends the Secure Storage Component the passcode entropy. The Secure Storage
Component first increments the counter for the lockbox. If the incremented counter
exceeds the maximum attempt value, the Secure Storage Component completely erases
the counter lockbox. If the maximum attempt count hasn’t been reached, the Secure
Storage Component attempts to derive the passcode verifier value and lockbox entropy
value with the same algorithm used to create the counter lockbox. If the derived passcode
verifier value matches the stored passcode verifier value, the Secure Storage Component
returns the lockbox entropy value to the Secure Enclave and resets the counter to 0.
The keys used to access password-protected data are rooted in the entropy stored in
counter lockboxes. For more information, see Data Protection overview.
The secure nonvolatile storage is used for all anti-replay services in the Secure Enclave.
Anti-replay services on the Secure Enclave are used for revocation of data over events that
mark anti-replay boundaries including, but not limited to, the following:
• Passcode change
• Enabling or disabling Touch ID or Face ID
• Adding or removing a Touch ID fingerprint or Face ID face
• Touch ID or Face ID reset
• Adding or removing an Apple Pay card
• Erase All Content and Settings
On architectures that don’t feature a Secure Storage Component, EEPROM (electrically

erasable programmable read-only memory) is utilized to provide secure storage services
for the Secure Enclave. Just like the Secure Storage Components, the EEPROM is attached
and accessible only from the Secure Enclave, but it doesn’t contain dedicated hardware
security features nor does it guarantee exclusive access to entropy (aside from its physical
attachment characteristics) nor counter lockbox functionality.

Secure Neural Engine
On devices with Face ID, the Secure Neural Engine converts 2D images and depth maps
into a mathematical representation of a user’s face.
On A11 through A13 SoCs, the Secure Neural Engine is integrated into the Secure Enclave.
The Secure Neural Engine uses direct memory access (DMA) for high performance. An
input-output memory management unit (IOMMU) under the sepOS kernel’s control limits
this direct access to authorized memory regions.
Starting with A14 and the M1, the Secure Neural Engine is implemented as a secure mode
in the Application Processor’s Neural Engine. A dedicated hardware security controller
switches between Application Processor and Secure Enclave tasks, resetting Neural Engine
state on each transition to keep Face ID data secure. A dedicated engine applies memory
encryption, authentication, and access control. At the same time, it uses a separate
cryptographic key and memory range to limit the Secure Neural Engine to authorized
memory regions.
Power and clock monitors

All electronics are designed to operate within a limited voltage and frequency envelope.
When operated outside this envelope, the electronics can malfunction and then security
controls may be bypassed. To ensure that the voltage and frequency stay in a safe
range, the Secure Enclave is designed with monitoring circuits. These monitoring circuits
are designed to have a much larger operating envelope than the rest of the Secure
Enclave. If the monitors detect an illegal operating point, the clocks in the Secure Enclave
automatically stop and don’t restart until the next SoC reset.
Secure Enclave feature summary

Note: A12, A13, S4, and S5 products first released in Fall 2020 have a 2nd-generation
Secure Storage Component; while earlier products based on these SoCs have 1st-
generation Secure Storage Component.
SoC Memory Protection Secure Storage AES Engine PKA

Engine
A8 Encryption and EEPROM Yes No

authentication
A9 Encryption and EEPROM DPA protection Yes

authentication
A10 Encryption, EEPROM DPA protection and OS-bound keys

authentication, and lockable seed bits
replay prevention
A11 Encryption, EEPROM DPA protection and OS-bound keys

authentication, and lockable seed bits
replay prevention
A12 (Apple devices Encryption, Secure Storage DPA protection and OS-bound keys
released before Fall authentication, and Component gen 1 lockable seed bits
2020) replay prevention

SoC Memory Protection Secure Storage AES Engine PKA
Engine
A12 (Apple devices Encryption, Secure Storage DPA protection and OS-bound keys
released after Fall authentication, and Component gen 2 lockable seed bits
A13 (Apple devices Encryption, Secure Storage DPA protection and OS-bound keys and
released before Fall authentication, and Component gen 1 lockable seed bits Boot Monitor
A13 (Apple devices Encryption, Secure Storage DPA protection and OS-bound keys and
released after Fall authentication, and Component gen 2 lockable seed bits Boot Monitor
A14 Encryption, Secure Storage DPA protection and OS-bound keys and
authentication, and Component gen 2 lockable seed bits Boot Monitor
replay prevention
S3 Encryption and EEPROM DPA protection and Yes

authentication lockable seed bits
S4 Encryption, Secure Storage DPA protection and OS-bound keys

authentication, and Component gen 1 lockable seed bits
replay prevention
S5 (Apple devices Encryption, Secure Storage DPA protection and OS-bound keys
released before Fall authentication, and Component gen 1 lockable seed bits
S5 (Apple devices Encryption, Secure Storage DPA protection and OS-bound keys
released after Fall authentication, and Component gen 2 lockable seed bits
S6 Encryption, Secure Storage DPA protection and OS-bound keys

authentication, and Component gen 2 lockable seed bits
replay prevention
T2 Encryption and EEPROM DPA protection and OS-bound keys

authentication lockable seed bits
M1 Encryption, Secure Storage DPA protection and OS-bound keys and

authentication, and Component gen 2 lockable seed bits Boot Monitor
replay prevention
Touch ID and Face ID

Touch ID and Face ID security
Passcodes and passwords are essential to the security of Apple devices. At the same time,
users need to be able to quickly access their devices, even exceeding a hundred times a
day. Biometric authentication provides a way to retain the security of a strong passcode—
or even strengthen the passcode or password because it won’t often need to be entered
manually—while providing the convenience of swiftly unlocking with a finger press or
glance. Touch ID and Face ID don’t replace a passcode or password, but in most situations
they do make access faster and easier.

Touch ID security
Touch ID is the fingerprint sensing system that makes secure access to supported Apple
devices faster and easier. This technology reads fingerprint data from any angle and
learns more about a user’s fingerprint over time, with the sensor continuing to expand the
fingerprint map as additional overlapping nodes are identified with each use.
Apple devices with a Touch ID sensor can be unlocked using a fingerprint. Touch ID doesn’t
replace the need for a device passcode or user password, which is still required after
device startup, restart, or logout (on a Mac). In some apps, Touch ID can also be used in
place of device passcode or user password—for example, to unlock password protected
notes in the Notes app, to unlock keychain-protected websites, and to unlock supported
app passwords. However, a device passcode or user password is always required in some
scenarios (for example, to change an existing device passcode or user password or to
remove existing fingerprint enrollments or create new ones).
When the fingerprint sensor detects the touch of a finger, it triggers the advanced imaging
array to scan the finger and sends the scan to the Secure Enclave. Communication
between the processor and the Touch ID sensor takes place over a serial peripheral
interface bus. The processor forwards the data to the Secure Enclave but can’t read it.
It’s encrypted and authenticated with a session key that’s negotiated using a shared key
provisioned for each Touch ID sensor and its corresponding Secure Enclave at the factory.
For every Touch ID sensor the shared key is strong, random, and different. The session key
exchange uses AES key wrapping, with both sides providing a random key that establishes
the session key and uses transport encryption that provides both authentication and
confidentiality (using AES-CCM).
While the fingerprint scan is being vectorized for analysis, the raster scan is temporarily
stored in encrypted memory within the Secure Enclave and then it’s discarded. The
analysis utilizes subdermal ridge flow angle mapping, a lossy process that discards “finger
minutiae data” that would be required to reconstruct the user’s actual fingerprint. The
resulting map of nodes is stored without any identity information in an encrypted format
that can be read only by the Secure Enclave. This data never leaves the device. It’s not
sent to Apple, nor is it included in device backups.
Face ID security
With a simple glance, Face ID securely unlocks supported Apple devices. It provides
intuitive and secure authentication enabled by the TrueDepth camera system, which uses
advanced technologies to accurately map the geometry of a user’s face. Face ID uses
neural networks for determining attention, matching, and antispoofing, so a user can
unlock their phone with a glance. Face ID automatically adapts to changes in appearance,
and carefully safeguards the privacy and security of a user’s biometric data.
Face ID is designed to confirm user attention, provide robust authentication with a low
false-match rate, and mitigate both digital and physical spoofing.
The TrueDepth camera automatically looks for the user’s face when the user wakes an
Apple device that features Face ID (by raising it or tapping the screen), as well as when
those devices attempt to authenticate the user in order to display an incoming notification
or when a supported app requests Face ID authentication. When a face is detected,
Face ID confirms attention and intent to unlock by detecting that the user’s eyes are open
and their attention is directed at their device; for accessibility, the Face ID attention check
is disabled when VoiceOver is activated and, if required, can be disabled separately.

After the TrueDepth camera confirms the presence of an attentive face, it projects and
reads over 30,000 infrared dots to form a depth map of the face along with a 2D infrared
image. This data is used to create a sequence of 2D images and depth maps, which are
digitally signed and sent to the Secure Enclave. To counter both digital and physical
spoofs, the TrueDepth camera randomizes the sequence of 2D images and depth map
captures, and projects a device-specific random pattern. A portion of the Secure Neural
Engine—protected within the Secure Enclave—transforms this data into a mathematical
representation and compares that representation to the enrolled facial data. This enrolled
facial data is itself a mathematical representation of the user’s face captured across a
variety of poses.
Touch ID, Face ID, passcodes, and passwords

To use Touch ID or Face ID, the user must set up their device so that a passcode or
password is required to unlock it. When Touch ID or Face ID detects a successful match,
the user’s device unlocks without asking for the device passcode or password. This makes
using a longer, more complex passcode or password far more practical because the user
doesn’t need to enter it as frequently. Touch ID and Face ID don’t replace the user’s
passcode or password; instead, they provide easy access to the device within thoughtful
boundaries and time constraints. This is important because a strong passcode or password
forms the foundation for how a user’s iPhone, iPad, Mac, or Apple Watch cryptographically
protects that user’s data.
When a device passcode or password is required

Users can use their passcode or password anytime instead of Touch ID or Face ID, but
there are situations where biometrics aren’t permitted. The following security-sensitive
operations always require entry of a passcode or password:
• Updating the software
• Erasing the device
• Viewing or changing passcode settings
• Installing configuration profiles
• Unlocking the Security & Privacy pane in System Preferences on Mac
• Unlocking the Users & Groups pane in System Preferences on Mac (if FileVault is turned
on)
A passcode or password is also required if the device is in the following states:
• The device has just been turned on or restarted.
• The user has logged out of their Mac account (or hasn’t yet logged in).
• The user hasn’t unlocked their device for more than 48 hours.
• The user hasn’t used their passcode or password to unlock their device for 156 hours
(six and a half days), and the user hasn’t used a biometric to unlock their device in 4
hours.
• The device has received a remote lock command.
• The user exited power off/Emergency SOS by pressing and holding either volume button
and the Sleep/Wake button simultaneously for 2 seconds and then pressing Cancel.

• There were five unsuccessful biometric match attempts (though for usability, the device
might offer entering a passcode or password instead of using biometrics after a smaller
number of failures).
When Touch ID or Face ID is enabled on an iPhone or iPad, the device immediately locks
when the Sleep/Wake button is pressed, and the device locks every time it goes to sleep.
Touch ID and Face ID require a successful match—or optionally the passcode—at every
wake.
The probability that a random person in the population could unlock a user’s iPhone,
iPad, or Mac is 1 in 50,000 with Touch ID or 1 in 1,000,000 with Face ID. This probability
increases with multiple enrolled fingerprints (up to 1 in 10,000 with five fingerprints) or
appearances (up to 1 in 500,000 with two appearances). For additional protection, both
Touch ID and Face ID allow only five unsuccessful match attempts before a passcode
or password is required to obtain access to the user’s device or account. With Face ID,
the probability of a false match is different for twins and siblings who look like the user
and for children under the age of 13 (because their distinct facial features may not have
fully developed). If a user is concerned about a false match, Apple recommends using a
passcode to authenticate.
Facial matching security

Facial matching is performed within the Secure Enclave using neural networks trained
specifically for that purpose. Apple developed the facial matching neural networks
using over a billion images, including infrared (IR) and depth images collected in studies
conducted with the participants’ informed consent. Apple then worked with participants
from around the world to include a representative group of people accounting for gender,
age, ethnicity, and other factors. The studies were augmented as needed to provide a
high degree of accuracy for a diverse range of users. Face ID is designed to work with
hats, scarves, eyeglasses, contact lenses, and many types of sunglasses. Furthermore,
it’s designed to work indoors, outdoors, and even in total darkness. An additional neural
network—that’s trained to spot and resist spoofing—defends against attempts to unlock
the device with photos or masks. Face ID data, including mathematical representations
of a user’s face, is encrypted and available only to the Secure Enclave. This data never
leaves the device. It’s not sent to Apple, nor is it included in device backups. The following
Face ID data is saved, encrypted only for use by the Secure Enclave, during normal
operation:
• The mathematical representations of a user’s face calculated during enrollment
• The mathematical representations of a user’s face calculated during some unlock

attempts if Face ID deems them useful to augment future matching
Face images captured during normal operation aren’t saved but are instead immediately
discarded after the mathematical representation is calculated for either enrollment or
comparison to the enrolled Face ID data.

Improving Face ID matches
To improve match performance and keep pace with the natural changes of a face and look,
Face ID augments its stored mathematical representation over time. Upon a successful
match, Face ID may use the newly calculated mathematical representation—if its quality
is sufficient—for a finite number of additional matches before that data is discarded.
Conversely, if Face ID fails to recognize a face but the match quality is higher than a
certain threshold and a user immediately follows the failure by entering their passcode,
Face ID takes another capture and augments its enrolled Face ID data with the newly
calculated mathematical representation. This new Face ID data is discarded if the user
stops matching against it or after a finite number of matches. These augmentation
processes allow Face ID to keep up with dramatic changes in a user’s facial hair or makeup
use while minimizing false acceptance.
Uses for Touch ID and Face ID
Unlocking a device or user account

With Touch ID or Face ID disabled, when a device or account locks, the keys for the highest
class of Data Protection—which are held in the Secure Enclave—are discarded. The files
and keychain items in that class are inaccessible until the user unlocks the device or
account by entering their passcode or password.
With Touch ID or Face ID enabled, the keys aren’t discarded when the device or account
locks; instead, they’re wrapped with a key that’s given to the Touch ID or Face ID
subsystem inside the Secure Enclave. When a user attempts to unlock the device or
account, if the device detects a successful match, it provides the key for unwrapping
the Data Protection keys, and the device or account is unlocked. This process provides
additional protection by requiring cooperation between the Data Protection and Touch ID or
Face ID subsystems to unlock the device.
When the device restarts, the keys required for Touch ID or Face ID to unlock the device or
account are lost; they’re discarded by the Secure Enclave after any condition is met that
requires passcode or password entry.
Securing purchases with Apple Pay

The user can also use Touch ID and Face ID with Apple Pay to make easy and secure
purchases in stores, apps, and on the web:
• Using Touch ID: For Touch ID, the intent to pay is confirmed using the gesture of
activating the Touch ID sensor combined with successfully matching the user’s
fingerprint.
• Using Face ID in stores: To authorize an in-store payment with Face ID, the user
must first confirm intent to pay by double-clicking the side button. This double-click
captures user intent using a physical gesture directly linked to the Secure Enclave
and is resistant to forgery by a malicious process. The user then authenticates using
Face ID before placing the device near the contactless payment reader. A different
Apple Pay payment method can be selected after Face ID authentication, which requires
reauthentication, but the user won’t have to double-click the side button again.

• Using Face ID in apps and on the web: To make a payment within apps and on the
web, the user confirms their intent to pay by double-clicking the side button and then
authenticates using Face ID to authorize the payment. If the Apple Pay transaction
isn’t completed within 60 seconds of double-clicking the side button, the user must
reconfirm intent to pay by double-clicking again.
Using system-provided APIs

Third-party apps can use system-provided APIs to ask the user to authenticate using
Touch ID or Face ID or a passcode or password, and apps that support Touch ID
automatically support Face ID without any changes. When using Touch ID or Face ID,
the app is notified only as to whether the authentication was successful; it can’t access
Touch ID, Face ID, or the data associated with the enrolled user.
Protecting keychain items

Keychain items can also be protected with Touch ID or Face ID, to be released by the
Secure Enclave only by a successful match or with the device passcode or account
password. App developers have APIs to verify that a passcode or password has been set
by the user before requiring Touch ID or Face ID or a passcode or password to unlock
keychain items. App developers can do any of the following:
• Require that authentication API operations don’t fall back to an app password or the
device passcode. They can query whether a user is enrolled, allowing Touch ID or
Face ID to be used as a second factor in security-sensitive apps.
• Generate and use Elliptic Curve Cryptography (ECC) keys inside the Secure Enclave
that can be protected by Touch ID or Face ID. Operations with these keys are always
performed inside the Secure Enclave after it authorizes their use.
Making and approving purchases

Users can also configure Touch ID or Face ID to approve purchases from the iTunes
Store, the App Store, Apple Books, and more, so users donʼt have to enter their
Apple ID password. When purchases are made, the Secure Enclave verifies that a biometric
authorization occurred and then releases ECC keys used to sign the store request.

Hardware microphone disconnect
All Apple silicon-based Mac notebooks and Intel-based Mac notebooks with the Apple T2
Security Chip feature a hardware disconnect that disables the microphone whenever the
lid is closed. On all 13-inch MacBook Pro and MacBook Air notebooks with the T2 chip,
all MacBook notebooks with a T2 chip from 2019 or later, and Mac notebooks with Apple
silicon, this disconnect is implemented in hardware alone. The disconnect prevents any
software—even with root or kernel privileges in macOS, and even the software on the T2
chip or other firmware—from engaging the microphone when the lid is closed. (The camera
isn’t disconnected in hardware, because its field of view is completely obstructed with the
lid closed.)
iPad models beginning in 2020 also feature the hardware microphone disconnect. When an
MFi-compliant case (including those sold by Apple) is attached to the iPad and closed, the
microphone is disconnected in hardware, preventing microphone audio data being made
available to any software—even with root or kernel privileges in iPadOS, or any device
firmware.
The protections in this section are implemented directly with hardware logic, according to
the following circuit diagram:
Circuit diagram.
In each product with a hardware microphone cutoff, one or more lid sensors detect the
physical closure of the lid or case using some physical property (for example, a Hall
effect sensor or a hinge angle sensor) of the interaction. For sensors where calibration is
necessary, parameters are set during production of the device and the calibration process
includes a nonreversible hardware lock out of any subsequent changes to sensitive
parameters on the sensor. These sensors emit a direct hardware signal that goes through a
simple set of nonreprogrammable hardware logic. This logic provides debounce, hysteresis,
and/or a delay of up to 500 ms before disabling the microphone. Depending on the product,
this signal can be implemented either by disabling the lines transporting data between
the microphone and the System on Chip (SoC) or by disabling one of the input lines to the
microphone module that’s allowing it to be active—for example, the clock line or a similar
effective control.

Express Cards with power reserve
If iOS isn’t running because iPhone needs to be charged, there may still be enough
power in the battery to support Express Card transactions. Supported iPhone devices
automatically support this feature with:
• A payment or transit card designated as the Express Transit card
• Student ID cards with Express Mode turned on
Pressing the side button (or on iPhone SE 2nd generation, the Home button) displays the
low-battery icon as well as text indicating that Express Cards are available to use. The NFC
controller performs Express Card transactions under the same conditions as when iOS
is running, except that transactions are indicated only with haptic notification (no visible
notification is shown). On iPhone SE 2nd generation, completed transactions may take
a few seconds to appear on screen. This feature isn’t available when a standard user-
initiated shutdown is performed.

System security
System security overview

Building on the unique capabilities of Apple hardware, system security is responsible for
controlling access to system resources in Apple devices without compromising usability.
System security encompasses the boot-up process, software updates, and protection of
computer system resources such as CPU, memory, disk, software programs, and stored
data.
The most recent versions of Apple operating systems are the most secure. An important
part of Apple security is secure boot, which protects the system from malware infection
at boot time. Secure boot begins in hardware and builds a chain of trust through software,
where each step ensures that the next is functioning properly before handing over control.
This security model supports not only the default boot of Apple devices but also the
various modes for recovery and timely updates on Apple devices. Subcomponents like the
T2 Chip and the Secure Enclave also perform their own secure boot to help ensure they
only boot known-good code from Apple. The update system can even prevent downgrade
attacks, so that devices can’t be rolled back to an older version of the operating system
(which an attacker knows how to compromise) as a method of stealing user data.
Apple devices also include boot and runtime protections so that they maintain their
integrity during ongoing operation. Apple-designed silicon on iPhone, iPad, Apple Watch,
Apple TV, HomePod, and a Mac with Apple silicon provide a common architecture for
protecting operating system integrity. macOS also features an expanded and configurable
set of protection capabilities in support of its differing computing model, as well as
capabilities supported on all Mac hardware platforms.
Secure boot
Boot process for iOS and iPadOS devices
Each step of the startup process contains components that are cryptographically signed
by Apple to enable integrity checking so that boot proceeds only after verifying the chain
of trust. These components include the bootloaders, the kernel, kernel extensions, and
cellular baseband firmware. This secure boot chain is designed to verify that the lowest
levels of software aren’t tampered with.

When an iOS or iPadOS device is turned on, its Application Processor immediately
executes code from read-only memory referred to as Boot ROM. This immutable code,
known as the hardware root of trust, is laid down during chip fabrication and is implicitly
trusted. The Boot ROM code contains the Apple Root certificate authority (CA) public key—
used to verify that the iBoot bootloader is signed by Apple before allowing it to load. This
is the first step in the chain of trust, in which each step checks that the next is signed by
Apple. When the iBoot finishes its tasks, it verifies and runs the iOS or iPadOS kernel. For
devices with an A9 or earlier A-series processor, an additional Low-Level Bootloader (LLB)
stage is loaded and verified by the Boot ROM and in turn loads and verifies iBoot.
A failure to load or verify following stages is handled differently depending on the

hardware:
• Boot ROM can’t load LLB (older devices): Device Firmware Upgrade (DFU) mode
• LLB or iBoot: Recovery mode
In either case, the device must be connected to iTunes (in macOS 10.14 or earlier) or the
Finder (macOS 10.15 or later) through USB and restored to factory default settings.
The Boot Progress Register (BPR) is used by the Secure Enclave to limit access to user
data in different modes and is updated before entering the following modes:
• DFU mode: Set by Boot ROM on devices with an Apple A12 or later SoCs
• Recovery mode: Set by iBoot on devices with Apple A10, S2, or later SoCs
On devices with cellular access, a cellular baseband subsystem performs additional secure
booting using signed software and keys verified by the baseband processor.
The Secure Enclave also performs a secure boot that checks its software (sepOS) is
verified and signed by Apple.
Memory safe iBoot implementation

In iOS 14 and iPadOS 14, Apple modified the C compiler toolchain used to build the iBoot
bootloader to improve its security. The modified toolchain implements code to prevent
memory- and type-safety issues that are typically encountered in C programs. For example,
it prevents:
• Buffer overflows, by ensuring that all pointers carry bounds information that is verified
when accessing memory
• Heap exploitation, by separating heap data from its metadata and accurately detecting
error conditions such as double free errors
• Type confusion, by ensuring that all pointers carry runtime type information that’s
verified during pointer cast operations
• Type confusion caused by use after free errors, by segregating all dynamic memory
allocations by static type
This technology is available on iPhone with Apple A13 Bionic or later, and iPad with the A14
Bionic chip.

Mac computers with Apple silicon
Boot process for a Mac with Apple silicon

When a Mac with Apple silicon is turned on, it performs a boot process much like that of
iPhone and iPad.
Boot process steps when a Mac with Apple silicon is started.
The chip executes code from the Boot ROM in the first step in the chain of trust. macOS
secure boot on a Mac with Apple silicon verifies not only the operating system code itself,
but also the security policies and even kexts (supported, though not recommended)
configured by authorized users.
When LLB is launched, it then verifies the signatures and loads system-paired firmware
for intra-SoC cores such as the storage, display, system management, and Thunderbolt
controllers. LLB is also responsible for loading the LocalPolicy, which is a file signed by
the Secure Enclave Processor. The LocalPolicy file describes the configuration that the
user has chosen for the system boot and runtime security policies. The LocalPolicy has the
same data structure format as all other boot objects, but it’s signed locally by a private key
that’s available only within a particular machine’s Secure Enclave instead of being signed
by a central Apple server (like software updates).
To prevent replay of any previous LocalPolicy, LLB must look up a nonce from the Secure
Enclave-attached Secure Storage Component. To do this, it uses the Secure Enclave
Boot ROM and makes sure the nonce in the LocalPolicy matches the nonce in the Secure
Storage Component. This prevents an old LocalPolicy—which could have been configured
for lower security—from being reapplied to the system after security has been upgraded.
The result is that secure boot on a Mac with Apple silicon helps protect not only against
rollback of operating system versions but also against security policy downgrades.

The LocalPolicy file captures whether the operating system is configured for Full, Reduced,
or Permissive security.
• Full Security: The system behaves like iOS and iPadOS, and allows only booting
software which was known to be the latest that was available at install time.
• Reduced Security: LLB is directed to trust “global” signatures, which are bundled with
the operating system. This allows the system to run older versions of macOS. Because
older versions of macOS inevitably have unpatched vulnerabilities, this security mode
is described as Reduced. This is also the policy level required to support booting kernel
extensions (kexts).
• Permissive Security: The system behaves like Reduced Security in that it uses global
signature verification for iBoot and beyond, but it also tells iBoot that it should accept
some boot objects being signed by the Secure Enclave with the same key used to sign
the LocalPolicy. This policy level supports users that are building, signing, and booting
their own custom XNU kernels.
If the LocalPolicy indicates to LLB that the selected operating system is running in Full
Security, LLB evaluates the personalized signature for iBoot. If it’s running in Reduced
Security or Permissive Security, it evaluates the global signature. Any signature verification
errors cause the system to boot to recoveryOS to provide repair options.
After LLB hands off to iBoot, it loads macOS-paired firmware such as that for the
Secure Neural Engine, the Always On Processor, and other firmware. iBoot also looks at
information about the LocalPolicy handed to it from LLB. If the LocalPolicy indicates that
there should be an Auxiliary Kernel Collection (AuxKC), iBoot looks for it on the file system,
verifies that it was signed by the Secure Enclave with the same key as the LocalPolicy, and
verifies that its hash matches a hash stored in the LocalPolicy. If the AuxKC is verified,
iBoot places it into memory with the Boot Kernel Collection before locking the full memory
region covering the Boot Kernel Collection and AuxKC with the System Coprocessor
Integrity Protection (SCIP). If the policy indicates that an AuxKC should be present but it
isn’t found, the system continues to boot into macOS without it. iBoot is also responsible
for verifying the root hash for the signed system volume (SSV), to check that the file
system the kernel will mount is fully integrity verified.

Boot modes for a Mac with Apple silicon
A Mac with Apple silicon has the boot modes described below.
Mode Key combo Description
macO S From a shutdown state, press and 1. Boot ROM hands off to LLB.
release the power button. 2. LLB loads system-paired firmware and the Local
Policy for the selected macO S.
3. LLB locks an indication into the Boot Progress
Register (BPR) that it’s booting into normal macOS,
and hands off to iB oot.
4. iB oot loads the macO S-paired firmware, the static
trust cache, the device tree, and the Boot Kernel
Collection.
5. If the LocalPolicy allows it, iB oot loads the Auxiliary
Kernel Collection (AuxKC) of third-party kexts.
6. If the LocalPolicy didn’t disable it, iB oot verifies the
root signature hash for the signed system volume
(SSV).
recoveryO S From a shutdown state, press and 1. Boot ROM hands off to LLB.
hold the power button. 2. LLB loads system-paired firmware and the Local
Policy for the recoveryO S.
3. LLB locks an indication into the Boot Progress
Register that it’s booting into recoveryO S, and hands
off to iB oot for recoveryO S.
4. iB oot loads the macO S-paired firmware, the
trust cache, the device tree, and the Boot Kernel
Collection.
Note: Security downgrades aren’t allowed on the
recoveryO S LocalPolicy.
Fallback recovery From a shutdown state, double- The same process as recoveryO S boot, except that
OS press and hold the power button. it boots to a second copy of recoveryO S that is kept
for resiliency. However, LLB doesn’t lock an indication
into the Boot Progress Register saying it is going
into recoveryO S, and therefore the fallback recovery
OS doesn’t have the capability to change the system
security state.
Safe mode Boot into recoveryO S per the above, 1. Boots to recoveryO S as per the above.
then hold Shift while selecting the 2. Holding the Shift key while selecting a volume
startup volume. causes the BootPicker application to approve that
macO S for booting, as normal, but to also set an
nvram variable that tells iB oot to not load the AuxKC
on the next boot.
3. System reboots and boots to the targeted volume,
but iB oot doesn’t load AuxKC.

Startup Disk security policy control for a Mac with Apple silicon
Overview
Unlike security policies on an Intel-based Mac, security policies on a Mac with Apple
silicon are for each installed operating system. This means that multiple installed macOS
instances with different versions and security policies are supported on the same machine.
For this reason, an operating system picker has been added to Startup Security Utility.
macOS storage selection to change the security policy.
On a Mac with Apple silicon, System Security Utility indicates the overall user-configured
security state of macOS, such as the booting of a kext or the configuration of System
Integrity Protection (SIP). If changing a security setting would significantly degrade
security or make the system easier to compromise, users must enter into recoveryOS by
holding the power button (so that malware can’t trigger the signal, only a human with
physical access can), in order to make the change. Because of this, an Apple-silicon based
Mac also won’t require (or support) a firmware password—all critical changes are already
gated by user authorization. For more information on SIP, see System Integrity Protection.
On a Mac with Apple silicon, System Security Utility indicates the overall user-configured
security state of macOS. And in many cases, if changing a security setting would
significantly degrade security or make the system easier to compromise, it’s a prerequisite
that a user enter into recoveryOS by holding the power button (so that malware can’t
trigger the signal, only a human with physical access can), in order to lower the operating
system security.
Full Security and Reduced Security can be set using Startup Security Utility from
recoveryOS. But Permissive Security can be accessed only from command-line tools for
users who accept the risk of making their Mac much less secure.

Full Security policy
Full Security is the default, and it behaves like iOS and iPadOS. At the time software is
downloaded and prepared to install, rather than using the global signature that comes with
the software, macOS contacts the same Apple signing server used for iOS and iPadOS and
requests a fresh, “personalized” signature. A signature is personalized when it includes the
Exclusive Chip Identification (ECID)—a unique ID specific to the Apple CPU in this case—as
part of the signing request. The signature given back by the signing server is then unique
and usable only by that particular Apple CPU. When the Full Security policy is in effect, the
Boot ROM and LLB ensures that a given signature isn’t just signed by Apple but is signed
for this specific Mac, essentially tying that version of macOS to that Mac.
macOS Full Security policy selection.
Using an online signing server also provides better protection against rollback attacks
than typical global signature approaches. In a global signing system, the security epoch
could have rolled many times, but a system that has never seen the latest firmware won’t
know this. For example, a computer that currently believes it’s in security epoch 1 accepts
software from security epoch 2, even if the current actual security epoch is 5. With an
Apple silicon online signing system, the signing server can reject creating signatures for
software that’s in anything except the latest security epoch.
Additionally, if an attacker discovers a vulnerability after a security epoch change, they

can’t simply pick up the vulnerable software from a previous epoch off system A and apply
it to system B in order to attack it. The fact that the vulnerable software from an older
epoch was personalized to system A prevents it from being transferable and thus being
used to attack a system B. All these mechanisms work together to provide much stronger
guarantees that attackers can’t purposely place vulnerable software on a Mac in order to
circumvent the protections provided by the latest software. But a user that’s in possession
of an administrator user name and password for the Mac can always choose the security
policy that works best for their use cases.

Reduced Security policy
Reduced Security is similar to Medium Security behavior on an Intel-based Mac with a
T2 chip, in which a vendor (in this case, Apple) generates a digital signature for the code
to assert it came from the vendor. In this way, attackers are prevented from inserting
unsigned code. Apple refers to this signature as a “global” signature because it can be
used on any Mac, for any amount of time, for a Mac that currently has a Reduced Security
policy set. Reduced security doesn’t itself provide protection against rollback attacks
(although unauthorized operating system changes can result in user data being rendered
inaccessible. For more information, see Kernel extensions in a Mac with Apple silicon.
macOS Reduced Security policy selection.
In addition to enabling users to run older versions of macOS, Reduced Security is required
for other actions that can put a user’s system security at risk, such as introducing third-
party kernel extensions (kexts). Kexts have the same privileges as the kernel, and thus
any vulnerabilities in third-party kexts can lead to full operating system compromise. This
is why developers are being strongly encouraged to adopt system extensions before kext
support is removed from macOS for future Mac computers with Apple silicon. Even when
third-party kexts are enabled, they can’t be loaded into the kernel on demand. Instead, the
kexts are merged into an Auxiliary Kernel Collection (AuxKC)—whose hash is stored in the
LocalPolicy—and thus they require a reboot. For more information about AuxKC generation,
see Kernel extensions in macOS.
Permissive Security policy

Permissive Security is for users who accept the risk of putting their Mac into a much more
insecure state. This mode differs from No Security mode on an Intel-based Mac with a T2
chip. With Permissive Security, signature verification is still performed along the entire
secure boot chain, but setting the policy to Permissive signals to iBoot that it should
accept locally Secure Enclave–signed boot objects, such as a user-generated Boot Kernel
Collection built from a custom XNU kernel. In this way, Permissive Security also provides
an architectural capability for running an arbitrary “fully untrusted operating system”
kernel. When a custom Boot Kernel Collection or fully untrusted operating system is loaded
on the system, some decryption keys become unavailable, to prevent a fully untrusted
operating systems from accessing data from trusted operating systems.

Important: Apple doesn’t provide or support custom XNU kernels.
macOS Permissive Security policy selection.
There’s another way that Permissive Security differs from No Security on an Intel-based
Mac with a T2 chip: It’s a prerequisite for some security downgrades that in the past have
been independently controllable. Most notably, to disable System Integrity Protection (SIP)
on a Mac with Apple silicon, a user must acknowledge that they’re putting the system into
Permissive Security. This is required because disabling SIP has always put the system into
a state that makes the kernel much easier to compromise. In particular, disabling SIP on a
Mac with Apple silicon disables kext signature enforcement during AuxKC generation time,
thus allowing any arbitrary kext to be loaded into kernel memory. Another improvement to
SIP that’s been made on a Mac with Apple silicon is that the policy store has been moved
out of NVRAM and into the LocalPolicy. So now, disabling SIP requires authentication by a
user who has access to the LocalPolicy signing key from recoveryOS (reached by pressing
and holding the power button). This makes it significantly more difficult for a software-only
attacker, or even a physically present attacker, to disable SIP.
It isn’t possible to downgrade to Permissive Security from the Startup Security Utility app.
Users can downgrade only by running command-line tools from Terminal in recoveryOS,
such as csrutil (to disable SIP). After the user has downgraded, the fact that it’s
occurred is reflected in Startup Security Utility, and so a user can easily set the security to
a more secure mode.
Note: A Mac with Apple silicon doesn’t require or support a specific media boot policy,
because technically all boots are performed locally. If a user chooses to boot from external
media, that operating system version must first be personalized using an authenticated
reboot from recoveryOS. This reboot creates a LocalPolicy file on the internal drive that’s
used to perform a trusted boot from the operating system stored on the external media.
This means the configuration of booting from external media is always explicitly enabled
on a per operating system basis, and already requires user authorization, so no additional
secure configuration is necessary.

LocalPolicy signing-key creation and management
Creation
When macOS is first installed in the factory, or when a tethered erase-install is performed,
the Mac runs code from temporary restore RAM disk to initialize the default state. During
this process, the restore environment creates a new pair of public and private keys
which are held in the Secure Enclave. The private key is referred to as the Owner Identity
Key (OIK). If any OIK already exists, it’s destroyed as part of this process. The restore
environment also initializes the key used for Activation Lock; the User Identity Key (UIK).
Part of that process which is unique to a Mac with Apple silicon is when UIK certification is
requested for Activation Lock, a set of requested constraints to be enforced at validation-
time on the LocalPolicy are included. If the device can’t get a UIK certified for Activation
Lock (for example, because the device is currently associated with a Find My Mac account
and reported as lost), its unable to proceed further to create a Local Policy. If a device
is issued a User identity Certificate (ucrt), that ucrt contains server imposed policy
constraints and user requested policy constraints in an X.509 v3 extension.
When an Activation Lock/ucrt is successfully retrieved, it’s stored in a database on the

server side and also returned to the device. After the device has a ucrt, a certification
request for the public key which corresponds to the OIK is sent to the Basic Attestation
Authority (BAA) server. BAA verifies the OIK certification request using the public key from
the ucrt stored in the BAA accessible database. If the BAA can verify the certification, it
certifies the public key, returning the Owner Identity Certificate (OIC) which is signed by
the BAA and contains the constraints stored in ucrt. The OIC is sent back to the Secure
Enclave. From then on, whenever the Secure Enclave signs a new LocalPolicy, it attaches
the OIC to the Image4. LLB has built-in trust in the BAA root certificate, which causes it to
trust the OIC, which causes it to trust the overall LocalPolicy signature.
RemotePolicy constraints
All Image4 files, not just Local Policies, contain constraints on Image4 manifest evaluation.
These constraints are encoded using special object identifiers (OIDs) in the leaf certificate.
The Image4 verification library looks up the special certificate constraint OID from a
certificate during signature evaluation and then mechanically evaluates the constraints
specified in it. The constraints are of the form:
• X must exist
• X must not exist
• X must have a specific value
So, for instance, for “personalized” signatures, the certificate constraints will contain
“ECID must exist,” and for “global” signatures, it will contain “ECID must not exist.” These
constraints ensure that all Image4 files signed by a given key must conform to certain
requirements to avoid erroneous signed Image4 manifest generation.

In the context of each LocalPolicy, these Image4 certificate constraints are referred to
as the RemotePolicy. A different RemotePolicy can exist for different boot environments’
LocalPolicies. The RemotePolicy is used to restrict the recoveryOS LocalPolicy so that
when recoveryOS is booted it can only ever behave as if it’s booting with Full Security.
This increases trust in the integrity of the recoveryOS boot environment as a place where
policy can be changed. The RemotePolicy restricts the LocalPolicy to contain the ECID
of the machine on which the LocalPolicy was generated, and the specific Remote Policy
Nonce Hash (rpnh) stored in the Secure Storage Component on that Mac. The rpnh, and
therefore the RemotePolicy, change only when actions are taken for Find My Mac and
Activation Lock, such as enrollment, unenrollment, remote lock, and remote wipe. Remote
Policy constraints are determined and specified at User Identity Key (UIK) certification
time and are signed in to the issued User identity Certificate (ucrt). Some Remote Policy
constraints are determined by the server such as ECID, ChipID and BoardID to prevent one
device from signing LocalPolicy files for another device. Other Remote Policy constraints
may be specified by the device to prevent Security downgrade of the Local Policy without
providing both the local authentication required to access the current OIK and remote
authentication of the account to which the device is Activation Locked.
Contents of a LocalPolicy file for a Mac with Apple silicon

The LocalPolicy is an Image4 file signed by the Secure Enclave. Image4 is an ASN.1
(Abstract Syntax Notation One) DER-encoded data structure format that’s used to describe
information about secure boot chain objects on Apple platforms. In an Image4-based
secure boot model, security policies are requested at software installation time initiated
by a signing request to a central Apple signing server. If the policy was acceptable, the
signing server returns a signed Image4 file containing a variety of four-character code
(4CC) sequences. These signed Image4 files and 4CCs are evaluated at startup by
software like the Boot ROM or LLB.
Ownership handoff between operating systems

Access to the Owner Identity Key (OIK) is referred to as “Ownership.” Ownership is required
to allow users to resign the LocalPolicy after making policy or software changes. The OIK
is protected with the same key hierarchy as described in Sealed Key Protection (SKP), with
the OIK being protected by the same Key encryption key (KEK) as the Volume encryption
key (VEK). This means it’s normally protected by both user passwords and measurements
of the operating system and policy. There’s only a single OIK for all operating systems on
the Mac. Therefore, when installing a second operating system, explicit consent is required
from the users on the first operating system to hand off Ownership to the users on the
second operating system. However, users don’t yet exist for the second operating system,
when the installer is running from the first operating system. Users in operating systems
aren’t normally generated until the operating system is booted and the Setup Assistant is
running. Thus two new actions are required when installing a second operating system on a
Mac with Apple silicon:
• Creating a LocalPolicy for the second operating system
• Preparing an “Install User” for handing off Ownership

When running an Install Assistant and targeting installation for a secondary blank volume,
a prompt asks the user if they’d like to copy a user from the current volume to be the
first user of the second volume. If the user says yes, the “Install User” which is created
is, in reality, a KEK which is derived from the selected user’s password and hardware keys,
which is then used to encrypt the OIK as it’s being handed to the second operating system.
Then from within the second operating system Install Assistant, it prompts for that user’s
password, to allow it to access the OIK in the Secure Enclave for the new operating system.
If users opt not to copy a user, the Install User is still created the same way, but an empty
password is used instead of a user’s password. This second flow exists for certain system
administration scenarios. However, users who want to have multi-volume installs and want
to perform Ownership handoff in the most secure fashion should always opt to copy a user
from the first operating system to the second operating system.
LocalPolicy on a Mac with Apple silicon

For a Mac with Apple silicon, local security policy control has been delegated to an
application running in the Secure Enclave. This software can utilize the user’s credentials
and the boot mode of the primary CPU to determine who can change the security policy
and from what boot environment. This helps prevent malicious software from using the
security policy controls against the user by downgrading them to gain more privileges.
LocalPolicy manifest properties

The LocalPolicy file contains some architectural 4CCs that are found in most all Image4
files—such as a board or model ID (BORD), indicating a particular Apple chip (CHIP), or
Exclusive Chip Identification (ECID). But the 4CCs below focus only on the security policies
that users can configure.
Note: Apple uses the term One True recoveryOS (1TR) to indicate a boot into the primary
recoveryOS which is achieved using a physical power button press. This is different from
a normal recoveryOS boot, which can be achieved using NVRAM or which may happen
when errors occur on startup. The physical button press increases trust that the boot
environment isn’t reachable by a software-only attacker who has broken into macOS.
LocalPolicy Nonce Hash (lpnh)

• Type: OctetString (48)
• Mutable environments: 1TR, recoveryOS, macOS

• Description: The lpnh is used for anti-replay of the LocalPolicy. This is an SHA384
hash of the LocalPolicy Nonce (LPN), which is stored in the Secure Storage Component
and accessible using the Secure Enclave Boot ROM or Secure Enclave. The raw nonce
is never visible to the Application Processor, only to the sepOS. An attacker wanting
to convince LLB that a previous LocalPolicy they had captured was valid would need
to place a value into the Secure Storage Component, which hashes to the same lpnh
value found in the LocalPolicy they want to replay. Normally there is a single LPN valid
on the system—except during software updates, when two are simultaneously valid—to
allow for the possibility of falling back to booting the old software in the event of an
update error. When any LocalPolicy for any operating system is changed, all policies are
re-signed with the new lpnh value corresponding to the new LPN found in the Secure
Storage Component. This change happens when the user changes security settings or
creates new operating systems with a new LocalPolicy for each.

Remote Policy Nonce Hash (rpnh)
• Description: The rpnh behaves the same way as the lpnh but is updated only when the
remote policy is updated, such as when changing the state of Find My enrollment. This
change happens when the user changes the state of Find My on their Mac.
recoveryOS Nonce Hash (ronh)

• Description: The ronh behaves the same way as the lpnh, but is found exclusively in
the LocalPolicy for recoveryOS. It’s updated when the recoveryOS is updated, such
as on software updates. A separate nonce from the lpnh and rpnh is used so that
when a device is put into a disabled state by Find My, existing operating systems can
be disabled (by removing their LPN and RPN from the Secure Storage Component),
while still leaving the recoveryOS bootable. In this way, the operating systems can be
reenabled when the system owner proves their control over the system by putting in
their iCloud password used for the Find My account. This change happens when a user
updates the recoveryOS or creates new operating systems.
Next Stage Image4 Manifest Hash (nsih)

• Description: The nsih field represents an SHA384 hash of the Image4 manifest data
structure that describes the booted macOS. The macOS Image4 manifest contains
measurements for all the boot objects—such as iBoot, the static trust cache, device
tree, Boot Kernel Collection, and signed system volume (SSV) volume root hash. When
LLB is directed to boot a given macOS, it ensures that the hash of the macOS Image4
manifest attached to iBoot matches what’s captured in the nsih field of the LocalPolicy.
In this way, the nsih captures the user intention of what operating system the user has
created a LocalPolicy for. Users change the nsih value implicitly when they perform a
software update.
Auxiliary Kernel Collection (AuxKC) Policy Hash (auxp)

• Mutable environments: macOS
• Description: The auxp is an SHA384 hash of the user-authorized kext list (UAKL) policy.
This is used at AuxKC generation time to ensure that only user-authorized kexts are
included in the AuxKC. smb2 is a prerequisite for setting this field. Users change the
auxp value implicitly when they change the UAKL by approving a kext from the Security
& Privacy pane in System Preferences.

Auxiliary Kernel Collection (AuxKC) Image4 Manifest Hash (auxi)
• Description: After the system verifies that the UAKL hash matches what’s found in
the auxp field of the LocalPolicy, it requests that the AuxKC be signed by the Secure
Enclave processor application that’s responsible for LocalPolicy signing. Next, an
SHA384 hash of the AuxKC Image4 manifest signature is placed into the LocalPolicy to
avoid the potential for mixing and matching previously signed AuxKCs to an operating
system at boot time. If iBoot finds the auxi field in the LocalPolicy, it attempts to load
the AuxKC from storage and validate its signature. It also verifies that the hash of the
Image4 manifest attached to the AuxKC matches the value found in the auxi field. If
the AuxKC fails to load for any reason, the system continues to boot without this boot
object and (so) without any third-party kexts loaded. The auxp field is a prerequisite
for setting the auxi field in the LocalPolicy. Users change the auxi value implicitly when
they change the UAKL by approving a kext from the Security & Privacy pane in System
Preferences.
Auxiliary Kernel Collection (AuxKC) Receipt Hash (auxr)

• Description: The auxr is an SHA384 hash of the AuxKC receipt, which indicates the
exact set of kexts that were included into the AuxKC. The AuxKC receipt can be a
subset of the UAKL, because kexts can be excluded from the AuxKC even if they’re
user authorized if they’re known to be used for attacks. In addition, some kexts that can
be used to break the user-kernel boundary may lead to decreased functionality, such
as an inability to use Apple Pay or play 4K and HDR content. Users who want these
capabilities opt in to a more restrictive AuxKC inclusion. The auxp field is a prerequisite
for setting the auxr field in the LocalPolicy. Users change the auxr value implicitly
when they build a new AuxKC from the Security & Privacy pane in System Preferences.
APFS volume group UUID (vuid)

• Description: The vuid indicates the volume group the kernel should use as root. This
field is primarily informational and isn’t used for security constraints. This vuid is set
by the user implicitly when creating a new operating system install.
Key encryption key (KEK) Group UUID (kuid)

• Description: The kuid indicates the volume that was booted. The key encryption key
has typically been used for Data Protection. For each LocalPolicy, it’s used to protect
the LocalPolicy signing key. The kuid is set by the user implicitly when creating a new
operating system install.

Paired recoveryOS Trusted Boot Policy Measurement (prot)
• Description: A paired recoveryOS Trusted Boot Policy Measurement (TBPM) is a special

iterative SHA384 hash calculation over the Image4 manifest of a LocalPolicy, excluding
nonces, in order to give a consistent measurement over time (because nonces like lpnh
are frequently updated). The prot field, which is found only in each macOS LocalPolicy,
provides a pairing to indicate the recoveryOS LocalPolicy that corresponds to the
macOS LocalPolicy.
Has Secure Enclave Signed recoveryOS Local Policy (hrlp)

• Type: Boolean
• Description: The hrlp indicates whether or not the prot value (above) is the
measurement of a Secure Enclave–signed recoveryOS LocalPolicy. If not, then the
recoveryOS LocalPolicy is signed by the Apple online signing server, which signs things
such as macOS Image4 files.
Secure Multi-Boot (smb0)

• Type: Boolean
• Mutable environments: 1TR, recoveryOS
• Description: If smb0 is present and true, LLB allows the next stage Image4 manifest to
be globally signed instead of requiring a personalized signature. Users can change this
field with Startup Security Utility or bputil to downgrade to Reduced Security.

• Type: Boolean
• Mutable environments: 1TR
• Description: If smb1 is present and true, iBoot allows objects such as a custom kernel
collection to be Secure Enclave signed with the same key as the LocalPolicy. Presence
of smb0 is a prerequisite for presence of smb1. Users can change this field using
command-line tools such as csrutil or bputil to downgrade to Permissive Security.

• Type: Boolean
• Description: If smb2 is present and true, iBoot allows the Auxiliary Kernel Collection to
be Secure Enclave signed with the same key as the LocalPolicy. The presence of smb0
is a prerequisite for the presence of smb2. Users can change this field using Startup
Security Utility or bputil to downgrade to Reduced Security and enable third-party
kexts.

• Type: Boolean
• Description: If smb3 is present and true, a user at the device has opted in to mobile
device management (MDM) control of their system. Presence of this field makes
the LocalPolicy-controlling Secure Enclave processor application accept MDM
authentication instead of requiring local user authentication. Users can change this field
using Startup Security Utility or bputil to enable managed control over third-party
kexts and software updates.

• Type: Boolean
• Mutable environments: recoveryOS, macOS
• Description: If smb4 is present and true, the device has opted in to MDM control of
the operating system using the Apple School Manager or Apple Business Manager.
Presence of this field makes the LocalPolicy-controlling Secure Enclave application
accept MDM authentication instead of requiring local user authentication. This field is
changed by the MDM solution when it detects that a device’s serial number appears in
Apple School Manager or Apple Business Manager.
System Integrity Protection (sip0)

• Type: 64 bit unsigned integer
• Description: The sip0 holds the existing System Integrity Protection (SIP) policy bits
that previously were stored in NVRAM. New SIP policy bits are added here (instead of
using LocalPolicy fields like the below) if they’re used only in macOS and not used by
LLB. Users can change this field using csrutil from 1TR to disable SIP and downgrade
to Permissive Security.

• Type: Boolean
• Description: If sip1 is present and true, iBoot should allow failures to verify the SSV
volume root hash. Users can change this field using csrutil or bputil from 1TR.

• Type: Boolean
• Description: sip2 allows disabling of the kernel System Coprocessor Integrity

Protection (SCIP). If this tag is present and true, iBoot tells the kernel that it should not
enforce SCIP. Users can change this field using csrutil or bputil from 1TR.

• Type: Boolean
• Description: If sip3 is present and true, iBoot doesn’t enforce its built-in allow list for
the boot-args NVRAM variable, which would otherwise filter the options passed to the
kernel. Users can change this field using csrutil or bputil from 1TR.
Certificates and RemotePolicy

As described in LocalPolicy signing-key creation and management, the LocalPolicy Image4
also contains the Owner Identity Certificate (OIC) and the embedded RemotePolicy.
Intel-based Mac computers
Boot process for an Intel-based Mac
Intel-based Mac with an Apple T2 Security Chip

When an Intel-based Mac computer with the Apple T2 Security Chip is turned on, the chip
performs a secure boot from its Boot ROM in the same fashion as iPhone, iPad, and a
Mac with Apple silicon. This verifies the iBoot bootloader and is the first step in the chain
of trust. iBoot checks the kernel and kernel extension code on the T2 chip, which then
checks the Intel UEFI firmware. The UEFI firmware and the associated signature are initially
available only to the T2 chip.
macOS T2 secure boot chain.

After verification, the UEFI firmware image is mapped into a portion of the T2 chip memory.
This memory is made available to the Intel CPU through the enhanced Serial Peripheral
Interface (eSPI). When the Intel CPU first boots, it fetches the UEFI firmware through the
eSPI from the integrity-checked, memory-mapped copy of the firmware located on the T2
chip.
The evaluation of the chain of trust continues on the Intel CPU, with the UEFI firmware
evaluating the signature for boot.efi, which is the macOS bootloader. The Intel-resident
macOS secure boot signatures are stored in the same Image4 format used for iOS, iPadOS,
and T2 chip secure boot, and the code that parses the Image4 files is the same hardened
code from the current iOS and iPadOS secure boot implementation. Boot.efi in turn verifies
the signature of a new file, called immutablekernel. When secure boot is enabled, the
immutablekernel file represents the complete set of Apple kernel extensions required to
boot macOS. The secure boot policy terminates at the handoff to the immutablekernel, and
after that, macOS security policies (such as System Integrity Protection and signed kernel
extensions) take effect.
If there are any errors or failures in this process, the Mac enters Recovery mode, Apple T2
Security Chip Recovery mode, or Apple T2 Security Chip Device Firmware Upgrade (DFU)
mode mode.
Microsoft Windows on an Intel-based Mac with a T2 chip

By default, an Intel-based Mac that supports secure boot trust only content signed by
Apple. However, to improve the security of Boot Camp installations, Apple also supports
secure booting for Windows. The UEFI firmware includes a copy of the Microsoft Windows
Production CA 2011 certificate used to authenticate Microsoft bootloaders.
Note: There is currently no trust provided for the Microsoft Corporation UEFI CA 2011 that
would allow verification of code signed by Microsoft partners. This UEFI CA is commonly
used to verify the authenticity of bootloaders for other operating systems, such as Linux
variants.
Support for secure boot of Windows isn’t enabled by default; instead, it’s enabled using
Boot Camp Assistant (BCA). When a user runs BCA, macOS is reconfigured to trust
Microsoft first-party signed code during boot. After BCA completes, if macOS fails to
pass the Apple first-party trust evaluation during secure boot, the UEFI firmware attempts
to evaluate the trust of the object according to UEFI secure boot formatting. If the trust
evaluation succeeds, the Mac proceeds and boots Windows. If not, the Mac enters
recoveryOS and informs the user of the trust evaluation failure.
Intel-based Mac computers without a T2 chip

An Intel-based Mac without a T2 chip doesn’t support secure boot. Therefore the UEFI
firmware loads the macOS booter (boot.efi) from the file system without verification, and
the booter loads the kernel (prelinkedkernel) from the file system without verification. To
protect the integrity of the boot chain, users should enable all of the following security
mechanisms:
• System Integrity Protection (SIP): Enabled by default, this protects the booter and
kernel against malicious writes from within a running macOS.
• FileVault: This can be enabled in two ways: by the user or by a mobile device
management (MDM) administrator. This protects against a physically present attacker
using Target Disk Mode to overwrite the booter.

• Firmware Password: This can be enabled in two ways: by the user or by an MDM
administrator. This protects a physically present attacker from launching alternate boot
modes such as recoveryOS, Single User Mode, or Target Disk Mode from which the
booter can be overwritten. This also prevents booting from alternate media, by which an
attacker could run code to overwrite the booter.
The unlocking process of an Intel-based Mac without a T2 chip.
Boot modes of an Intel-based Mac with an Apple T2 Security Chip

An Intel-based Mac with an Apple T2 Security Chip has a variety of boot modes that can
be entered at boot time by pressing key combinations, which are recognized by the UEFI
firmware or booter. Some boot modes, such as Single User Mode, won’t work unless the
security policy is changed to No Security in Startup Security Utility.
macO S boot None The UEFI firmware hands off to the

macO S booter (a UEFI application),
which hands off to the macOS
kernel. On standard booting of a
Mac with FileVault enabled, the
macO S booter presents the Login
Window interface, which takes the
password to decrypt the storage.
Startup Manager Option (⌥) The UEFI firmware launches the

built-in UEFI application that
presents the user with a boot device
selection interface.
Target Disk Mode (TDM) T The UEFI firmware launches the

built-in UEFI application that
exposes the internal storage device
as a raw, block-based storage
device over FireWire, Thunderbolt,
USB, or any combination of the
three (depending on the Mac
model).
Single User Mode Command (⌘)-S The macO S kernel passes the -s
flag in launchd’s argument vector,
then launchd creates the single-
user shell in the Console app’s tty.
Note: If the user exits the shell,

macO S continues boot to the Login
Window.
recoveryO S Command (⌘)-R The UEFI firmware loads a minimal

macO S from a signed disk image
(.dmg) file on the internal storage
device.
Internet recoveryO S Option (⌥)-Command (⌘)-R The signed disk image is

downloaded from the internet using
HTTP.

Diagnostics D The UEFI firmware loads a minimal

UEFI diagnostic environment from
a signed disk image file on the
internal storage device.
Internet Diagnostics Option (⌥)-D The signed disk image is

downloaded from the internet using
HTTP.
Windows boot None If Windows has been installed using

Boot Camp, the UEFI firmware
hands off to the Windows booter,
which hands off to the Windows
kernel.
Startup Security Utility on a Mac with an Apple T2 Security Chip
Overview
On an Intel-based Mac with an Apple T2 Security Chip, Startup Security Utility handles a
number security policy settings. The utility is accessible by booting into recoveryOS and
selecting Startup Security Utility from the Utilities menu and protects supported security
settings from easy manipulation by an attacker.
A screenshot of the Startup Security Utility.

Critical policy changes require authentication, even in Recovery mode. When Startup
Security Utility is first opened, it prompts the user to enter an administrator password
from the primary macOS installation associated with the currently booted recoveryOS. If
no administrator exists, one must be created before the policy can be changed. The T2
chip requires that the Mac computer is currently booted into recoveryOS and that an
authentication with a Secure Enclave–backed credential has occurred before such a policy
change can be made. Security policy changes have two implicit requirements. recoveryOS
must:
• Be booted from a storage device directly connected to the T2 chip, because partitions
on other devices don’t have Secure Enclave–backed credentials bound to the internal
storage device.
• Reside on an APFS-based volume, because there is support only for storing the
Authentication in Recovery credentials sent to the Secure Enclave on the “Preboot”
APFS volume of a drive. HFS plus-formatted volumes can’t use secure boot.
This policy is shown only in Startup Security Utility on an Intel-based Mac with a T2 chip.
Although most use cases shouldn’t require changes to the secure boot policy, users are
ultimately in control of their device’s settings and may choose, depending on their needs,
to disable or downgrade the secure boot functionality on their Mac.
Secure boot policy changes made from within this app apply only to the evaluation of the
chain of trust being verified on the Intel processor. The option “Secure boot the T2 chip” is
always in effect.
Secure boot policy can be configured to one of three settings: Full Security, Medium
Security, and No Security. No Security completely disables secure boot evaluation on the
Intel processor and allows the user to boot whatever they want.
Full Security boot policy

Full Security is the default boot policy, and it behaves similar to iOS and iPadOS or Full
Security on a Mac with Apple silicon. At the time that software is downloaded and prepared
to install, it is personalized with a signature that includes the Exclusive Chip Identification
(ECID)—a unique ID specific to the T2 chip in this case—as part of the signing request. The
signature given back by the signing server is then unique and usable only by that particular
T2 chip. When the Full Security policy is in effect, the UEFI firmware ensures that a given
signature isn’t just signed by Apple but is signed for this specific Mac, essentially tying
that version of macOS to that Mac. This helps prevent rollback attacks as described for
Full Security on a Mac with Apple silicon.
Medium Security boot policy

Medium Security boot policy is somewhat like a traditional UEFI secure boot, in which a
vendor (here, Apple) generates a digital signature for the code to assert it came from the
vendor. In this way, attackers are prevented from inserting unsigned code. We refer to
this signature as a “global” signature because it can be used on any Mac, for any amount
of time, for a Mac that currently has a Medium Security policy set. Neither iOS, iPadOS,
nor the T2 chip itself support global signatures. This setting doesn’t attempt to prevent
rollback attacks.

Media boot policy
Media boot policy exists only on an Intel-based Mac with a T2 chip and is independent
from the secure boot policy. So even if a user disables secure boot, this doesn’t change
the default behavior of preventing anything other than the storage device directly
connected to the T2 chip to boot the Mac. (Media boot policy is not required on a Mac with
Apple silicon. For more information, see Startup Disk security policy control.
Firmware password protection in an Intel-based Mac

macOS on Intel-based Mac computers with an Apple T2 Security Chip supports the use
of a Firmware Password to prevent unintended modifications of firmware settings on a
specific Mac. The Firmware Password is used to prevent selecting alternate boot modes
such as booting into recoveryOS or Single User Mode, booting from an unauthorized
volume, or booting into Target Disk Mode.
Note: The firmware password isn’t required on a Mac with Apple silicon, because the
critical firmware functionality it restricted has been moved into the recoveryOS and
(when FileVault is enabled) recoveryOS requires user authentication before its critical
functionality can be reached.
The most basic mode of firmware password can be reached from the recoveryOS Firmware
Password Utility on an Intel-based Mac without a T2 chip, and from the Startup Security
Utility on an Intel-based Mac with a T2 chip. Advanced options (such as the ability to
prompt for the password at every boot) are available from the firmwarepasswd command-
line tool in macOS.
Setting a Firmware Password is especially important to reduce the risk of attacks on Intel-
based Macs without a T2 chip from a physically present attacker. The Firmware Password
can help prevent an attacker from booting to recoveryOS, from where they could otherwise
disable System Integrity Protection (SIP). And by restricting boot of alternative media, an
attacker can’t execute privileged code from another operating system to attack peripheral
firmwares.
A firmware password reset mechanism exists to help users who forget their password.
Users press a key combination at startup, and are presented with a model-specific string
to provide to AppleCare. AppleCare digitally signs a resource that is signature checked by
the Uniform Resource Identifier (URI). If the signature is validated and the content is for
the specific Mac, the UEFI firmware removes the firmware password.
For users who want no one but themselves to remove their firmware password by software
means, the -disable-reset-capability option has been added to the firmwarepasswd
command-line tool in macOS 10.15. Before setting this option, users must acknowledge
that if the password is forgotten and needs removal, the user must bear the cost of the
logic board replacement necessary to achieve this. Organizations that want to protect their
Mac computers from external attackers and from employees must set a firmware password
on organization-owned systems. This can be accomplished on the device in any of the
following ways:
• At provisioning time, by manually using the firmwarepasswd command-line tool
• With third-party management tools that use the firmwarepasswd command-line tool
• Using mobile device management (MDM)

recoveryOS and diagnostics environments for an Intel-based Mac
recoveryOS
The recoveryOS is completely separate from the main macOS, and the entire contents
are stored in a disk image file named BaseSystem.dmg. There is also an associated
BaseSystem.chunklist which is used to verify the integrity of the BaseSystem.dmg. The
chunklist is a series of hashes for 10 MB chunks of the BaseSystem.dmg. The UEFI
firmware evaluates the signature of the chunklist file and then evaluates the hash for one
chunk at a time from the BaseSystem.dmg, to ensure that it matches the signed content
present in the chunklist. If any of these hashes don’t match, booting from the local
recoveryOS is aborted and the UEFI firmware attempts to boot from Internet recoveryOS
instead.
If the verification is successfully completed, the UEFI firmware mounts the BaseSystem.
dmg as a RAM disk and launches the boot.efi file that’s in it. There’s no need for the UEFI
firmware to do a specific check of the boot.efi, nor for the boot.efi to do a check of the
kernel, because the completed contents of the operating system (of which these elements
are only a subset) have already been integrity checked.
Apple Diagnostics
The procedure for booting the local diagnostic environment is mostly the same as
launching the recoveryOS. Separate AppleDiagnostics.dmg and AppleDiagnostics.chunklist
files are used, but they’re verified in the same way as the BaseSystem files are. Instead
of launching boot.efi, the UEFI firmware launches a file inside the disk image (.dmg file)
named diags.efi, which is in turn responsible for invoking a variety of other UEFI drivers
that can interface with and check for errors in the hardware.
Internet recoveryOS and diagnostic environment

If an error has occurred in the launching of the local recovery or diagnostic environments,
the UEFI firmware attempts to download the images from the internet instead. (A user
can also specifically request the images to be fetched from the internet using special
key sequences held at boot.) The integrity validation of the disk images and chunklists
downloaded from the OS Recovery Server is performed the same way as with images
retrieved from a storage device.
While the connection to the OS Recovery Server is done using HTTP, the complete
downloaded contents are still integrity checked as previously described, and as such are
protected against manipulation by an attacker with control of the network. In the event
that an individual chunk fails integrity verification, it is re-requested from the OS Recovery
Server 11 times, before giving up and displaying an error.
When the internet recovery and diagnostic modes were added to Mac computers in 2011,
it was decided that it would be better to use the simpler HTTP transport, and handle
content authentication using the chunklist mechanism, rather than implement the more
complicated HTTPS functionality in the UEFI firmware, and thus increase the firmwareʼs
attack surface.

Secure software updates
Overview
Security is a process; it isn’t enough to reliably boot the operating system version installed
at the factory—there must also exist a mechanism to quickly and securely obtain the latest
security updates. Apple regularly releases software updates to address emerging security
concerns. Users of iOS and iPadOS devices receive update notifications on the device and
through the Finder (in macOS 10.15 or later) and iTunes (in macOS 10.14 or earlier). Mac
users find available updates in System Preferences. Updates are delivered wirelessly, for
rapid adoption of the latest security fixes.
The update process uses the same hardware-based root of trust that secure boot uses
that is designed to install only Apple-signed code. The update process also uses system
software authorization to check that only copies of operating system versions that are
actively being signed by Apple can be installed on iOS and iPadOS devices, or on Mac
computers with the Full Security setting configured as the secure boot policy in Startup
Security Utility. With these secure processes in place, Apple can stop signing older
operating system versions with known vulnerabilities and help prevent downgrade attacks.
For greater software update security, when the device to be upgraded is physically
connected to a Mac, a full copy of iOS or iPadOS is downloaded and installed. But for over-
the-air (OTA) software updates, only the components required to complete an update are
downloaded, improving network efficiency by not downloading the entire operating system.
What’s more, software updates can be cached on a Mac running macOS 10.13 or later with
Content Caching turned on, so that iOS and iPadOS devices don’t need to redownload the
necessary update over the internet. (They still need to contact Apple servers to complete
the update process.)

Personalized update process
During upgrades and updates, a connection is made to the Apple installation authorization
server, which includes a list of cryptographic measurements for each part of the
installation bundle to be installed (for example, iBoot, the kernel, and the operating system
image), a random anti-replay value (the nonce), and the device’s unique Exclusive Chip
Identification (ECID).
The authorization server checks the presented list of measurements against versions for
which installation is permitted and, if it finds a match, adds the ECID to the measurement
and signs the result. The server passes a complete set of signed data to the device as
part of the upgrade process. Adding the ECID “personalizes” the authorization for the
requesting device. By authorizing and signing only for known measurements, the server
ensures that the update takes place exactly as Apple provided.
How Apple devices interact with the Apple Authorization server.
The boot-time chain-of-trust evaluation verifies that the signature comes from Apple and
that the measurement of the item loaded from the storage device, combined with the
device’s ECID, matches what was covered by the signature. These steps ensure that, on
devices that support personalization, the authorization is for a specific device and that
an older operating system of firmware version from one device can’t be copied to another.
The nonce prevents an attacker from saving the server’s response and using it to tamper
with a device or otherwise alter the system software.
The personalization process is why a network connection to Apple is always required to

update any device with Apple-designed silicon, including an Intel-based Mac with the
Apple T2 Security Chip.
Finally, the user’s data volume is never mounted during a software update, to help prevent
anything being read from or written to that volume during updates.
On devices with the Secure Enclave, that hardware similarly uses system software
authorization to check the integrity of its software and prevent downgrade installations.

Operating system integrity
Apple designs its operating systems with security at the core, with a hardware root of trust
leveraged to enable secure boot and a secure software update process that enables quick
and safe application of security updates. Apple operating systems also leverage purpose-
built silicon-based hardware capabilities to help prevent exploitation as the operating
system runs. These capabilities are not focused on whether or not to execute code (see
App security for the mechanisms that help prevent malware and ensure only trusted code
is executed) but instead focus on the capabilities afforded to code that is executed. In
short, they help mitigate attack and exploit techniques whether those originate from a
malicious app, from the web, or through any other channel. Protections listed here are
available on devices with supported Apple-designed SoCs—this includes iOS, iPadOS,
tvOS, watchOS, and now macOS on a Mac with Apple silicon.
Feature A10 A11, S3 A12, S4 A13, S5 A14, S6 M1
Kernel
Integrity
Protection
Fast
Permission
Restrictions
System
Coprocessor
Integrity
Protection
Pointer
Authentication
Codes
Page See Note

Protection below.
Layer
Note: Page Protection Layer (PPL) requires that the platform execute only signed and
trusted code; this is a security model that isn’t applicable on macOS.
Kernel Integrity Protection

After the operating system kernel completes initialization, Kernel Integrity Protection (KIP)
is enabled to prevent modifications of kernel and driver code. The memory controller
provides a protected physical memory region that iBoot uses to load the kernel and kernel
extensions. After startup completes, the memory controller denies writes to the protected
physical memory region. The Application Processor’s Memory Management Unit (MMU)
is configured to prevent mapping privileged code from physical memory outside the
protected memory region and to prevent writeable mappings of physical memory within the
kernel memory region.
To prevent reconfiguration, the hardware used to enable KIP is locked after the boot
process is complete.

Fast Permission Restrictions
Starting with the Apple A11 Bionic and S3 SoCs, a new hardware primitive was introduced.
This primitive, Fast Permission Restrictions, includes a CPU register that quickly restricts
permissions per thread. With Fast Permission Restrictions (also known as APRR registers),
supported operating systems can remove execute permissions from memory without the
overhead of a system call and a page table walk or flush. These registers provide one more
level of mitigation for attacks from the web, particularly for code compiled at runtime (just-
in-time compiled)—because memory can’t be effectively executed at the same time it’s
being read from and written to.
System Coprocessor Integrity Protection

Coprocessor firmware handles many critical system tasks—for example, the Secure
Enclave, the image sensor processor, and the motion coprocessor. Therefore its security
is a key part of the security of the overall system. To prevent modification of coprocessor
firmware, Apple uses a mechanism called System Coprocessor Integrity Protection (SCIP).
SCIP works much like Kernel Integrity Protection (KIP): At boot time, iBoot loads each
coprocessor’s firmware into a protected memory region, one that’s reserved and separate
from the KIP region. iBoot configures each coprocessor’s memory unit to prevent:
• Executable mappings outside its part of the protected memory region
• Writeable mappings inside its part of the protected memory region
Also at boot time, to configure SCIP for the Secure Enclave, the Secure Enclave operating
system is used. After the boot process is complete, the hardware used to enable SCIP is
locked to prevent reconfiguration.

Pointer Authentication Codes
Pointer Authentication Codes (PACs) are used to protect against exploitation of memory
corruption bugs. System software and built-in apps use PAC to prevent modification of
function pointers and return addresses (code pointers). PAC uses five secret 128-bit
values to sign kernel instructions and data, and each user space process has its own B
keys. Items are salted and signed as indicated below.
Item Key Salt
Function Return Address IB Storage address
Function Pointers IA 0
Block Invocation Function IA Storage address
Objective-C Method Cache IB Storage address + Class + Selector
C++ V-Table Entries IA Storage address + Hash (mangled

method name)
Computed Goto Label IA Hash (function name)
Kernel Thread State GA •
User Thread State Registers IA Storage address
C++ V-Table Pointers DA 0
The signature value is stored in the unused padding bits at the top of the 64-bit pointer.
The signature is verified before use, and the padding is restored to ensure a functioning
pointer address. Failure to verify results in an abort. This verification increases the
difficulty of many attacks, such as a return-oriented programming (ROP) attack, which
attempts to trick the device into executing existing code maliciously by manipulating
function return addresses stored on the stack.
Page Protection Layer

Page Protection Layer (PPL) in iOS, iPadOS, and watchOS prevents user space code from
being modified after code signature verification is complete. Building on Kernel Integrity
Protection and Fast Permission Restrictions, PPL manages the page table permission
overrides to make sure only the PPL can alter protected pages containing user code and
page tables. The system provides a massive reduction in attack surface by supporting
systemwide code integrity enforcement, even in the face of a compromised kernel.
This protection isn’t offered in macOS because PPL is only applicable on systems where
all executed code must be signed.

Additional macOS system security capabilities
Additional macOS system security capabilities
macOS operates on a broader set of hardware (for example, Intel-based CPUs, Intel-based
CPUs in combination with the Apple T2 Security Chip, and Apple silicon-based SoCs)
and supports a range of general-purpose computing use cases. Whereas some users use
only the basic preinstalled apps or those available from the App Store, others are kernel
hackers who need to disable essentially all platform protections so they can run and test
their executing code as with the highest levels of trust. Most fall somewhere between, and
many of those have peripherals and software that require varying levels of access. Apple
designed the macOS platform with an integrated approach to hardware, software, and
services—a platform that provides security by design and makes it simple to configure,
deploy, and manage but that retains the configurability that users expect. macOS also
includes the key security technologies that an IT professional needs to help protect
corporate data and integrate within secure enterprise networking environments.
The following capabilities support and help secure the varied needs of macOS users. They
include:
• Signed system volume security
• System Integrity Protection
• Trust caches
• Protection for peripherals
• Rosetta 2 (automatic translation) support and security for a Mac with Apple silicon
• DMA support and protections
• Kernel extension (kext) support and security
• Option ROM support and security
• UEFI firmware security for Intel-based Mac computers
Signed system volume security in macOS

For macOS 10.15, Apple introduced the read-only system volume, a dedicated, isolated
volume for system content. macOS 11 adds strong cryptographic protections to system
content with a signed system volume (SSV). SSV features a kernel mechanism that verifies
the integrity of the system content at runtime and rejects any data—code and noncode—
without a valid cryptographic signature from Apple.
SSV not only helps prevent tampering with any Apple software that’s part of the operating
system, it also makes macOS software update more reliable and much safer. And because
SSV uses APFS (Apple File System) snapshots, if an update can’t be performed, the old
system version can be restored without reinstallation.

Since its introduction, APFS has provided file-system metadata integrity using
noncryptographic checksums on the internal storage device. SSV strengthens the integrity
mechanism by adding cryptographic hashes, thus extending it to encompass every
byte of file data. Data from the internal storage device (including file system metadata)
is cryptographically hashed in the read path, and the hash is then compared with an
expected value in the file-system metadata. In case of mismatch, the system assumes the
data has been tampered with and won’t return it to the requesting software.
Each SSV SHA256 hash is stored in the main file-system metadata tree, which is itself
hashed. And because each node of the tree recursively verifies the integrity of the hashes
of its children—similar to a binary hash (Merkle) tree—the root node’s hash value, called a
seal, therefore encompasses every byte of data in the SSV, which means the cryptographic
signature covers the entire system volume.
During macOS installation and update, the seal is recomputed from the file system on-
device and that measurement is verified against the measurement which Apple signed. On
a Mac with Apple silicon, the bootloader verifies the seal before transferring control to the
kernel. On an Intel-based Mac with an Apple T2 Security Chip, the bootloader forwards
the measurement and signature to the kernel, which then verifies the seal directly before
mounting the root file system. In either case, if the verification fails, the startup process
halts and the user is prompted to reinstall macOS.
This procedure is repeated at every boot unless the user has elected to enter a lower
security mode and has separately chosen to disable the signed system volume.
SSV and code signing

Code signing is still present and enforced by the kernel. The signed system volume
provides protection when any bytes at all are read from the internal storage device. In
contrast, code signing provides protection when Mach objects are memory mapped as
executable. Both SSV and code signing protect executable code on all read and execute
paths.
SSV and FileVault

In macOS 11, equivalent at-rest protection for system content is provided by the SSV, and
therefore the system volume no longer needs to be encrypted. Any modifications made to
the file system while it’s at rest will be detected by the file system when they’re read. If the
user has enabled FileVault, the user’s content on the data volume is still encrypted with a
user-provided secret.
If the user chooses to disable the SSV, the system at rest becomes vulnerable to
tampering, and this tampering could enable an attacker to extract encrypted user data
when the system next starts up. Therefore the system won’t permit the user to disable the
SSV if FileVault is enabled. Protection while at rest must be enabled or disabled for both
volumes in a consistent manner.
In macOS 10.15 or earlier, FileVault protects operating system software while at rest by
encrypting user and system content with a key protected by a user-provided secret.
This protects against an attacker with physical access to the device from accessing or
effectively modifying the file system containing system software.

SSV and a Mac with an Apple T2 Security Chip
On a Mac with an Apple T2 Security Chip, only macOS itself is protected by the SSV. The
software that runs on the T2 chip and verifies macOS is protected by secure boot.
System Integrity Protection

macOS utilizes kernel permissions to limit writability of critical system files with a feature
called System Integrity Protection (SIP). This feature is separate and in addition to the
hardware-based Kernel Integrity Protection (KIP) available on a Mac with Apple silicon,
which protects modification of the kernel in memory. Mandatory access control technology
is leveraged to provide this and a number of other kernel level protections, including
sandboxing and Data Vault.
Mandatory access controls

macOS uses mandatory access controls—policies that set security restrictions, created
by the developer, that can’t be overridden. This approach is different from discretionary
access controls, which permit users to override security policies according to their
preferences.
Mandatory access controls aren’t visible to users, but they’re the underlying technology
that helps enable several important features, including sandboxing, parental controls,
managed preferences, extensions, and System Integrity Protection.
System Integrity Protection

System Integrity Protection restricts components to read-only in specific critical file
system locations to prevent malicious code from modifying them. System Integrity
Protection is a computer-specific setting that’s on by default when a user upgrades to
OS X 10.11 or later. On an Intel-based Mac, disabling it removes protection for all partitions
on the physical storage device. macOS applies this security policy to every process
running on the system, regardless of whether it’s running sandboxed or with administrative
privileges.
Trust caches
One of the objects included in the Secure Boot chain is the static trust cache, a trusted
record of all the Mach-O binaries that are mastered into the signed system volume. Each
Mach-O is represented by a code directory hash. For efficient searching, these hashes are
sorted before being inserted into the trust cache. The code directory is the result of the
signing operation performed by codesign(1). To enforce the trust cache, SIP must remain
enabled. To disable trust cache enforcement on a Mac with Apple silicon, secure boot must
be configured to Permissive Security.
When a binary is executed (whether as part of spawning a new process or mapping

executable code into an existing process), its code directory is extracted and hashed. If the
resulting hash is found in the trust cache, the executable mappings created for the binary
will be granted platform privileges—that is, they may possess any entitlement and execute
without further verification as to the authenticity of the signature. This is in contrast to
an Intel-based Mac, where platform privileges are conveyed to operating system content
by the Apple certificate that signs the binaries. (This certificate doesn’t constrain which
entitlements the binary may possess.)

Nonplatform binaries (for example, notarized third-party code) must have valid certificate
chains in order to execute, and the entitlements they may possess are constrained by the
signing profile issued to the developer by the Apple Developer Program.
All binaries shipped within macOS are signed with a platform identifier. On a Mac with
Apple silicon, this identifier is used to indicate that even though the binary is signed by
Apple, its code directory hash must be present in the trust cache in order to execute.
On an Intel-based Mac, the platform identifier is used to perform targeted revocation of
a binaries from an older release of macOS and prevent them from executing on newer
versions.
The static trust cache completely locks a set of binaries to a given version of macOS. This
behavior prevents legitimately Apple-signed binaries from older operating systems from
being introduced into newer ones in order for an attacker to gain advantage.
Platform code shipped outside the operating system

Apple ships some binaries—for example, Xcode and the development tools stack—that
aren’t signed with a platform identifier. Even so, they’re still permitted to execute with
platform privileges on a Mac with Apple silicon and those with a T2 chip. Because this
platform software is shipped independently of macOS, it isn’t subject to the revocation
behaviors imposed by the static trust cache.
Loadable trust caches

Apple ships certain software packages with loadable trust caches. These caches have the
same data structure as the static trust cache. But although there’s only one static trust
cache—and its contents are always guaranteed to be locked into read-only ranges after
the kernel’s early initialization is complete—loadable trust caches are added to the system
at runtime.
These trust caches are authenticated either through the same mechanism that
authenticates boot firmware (personalization using the Apple trusted signing service) or as
globally signed objects (whose signatures don’t bind them to a particular device).
One example of a personalized trust cache is the cache, shipped with the disk image
that’s used to perform field diagnostics on a Mac with Apple silicon. This trust cache is
personalized, along with the disk image, and loaded into the subject Mac computer’s
kernel while it’s booted into a diagnostic mode. The trust cache enables the software
within the disk image to run with platform privilege.
An example of a globally signed trust cache is shipped with macOS software updates. This
trust cache permits a chunk of code within the software update—the update brain—to run
with platform privilege. The update brain performs any work to stage the software update
that the host system lacks the capacity to perform in a consistent fashion across versions.

Peripheral processor security in Mac computers
All modern computing systems have many built-in peripheral processors dedicated to tasks
such as networking, graphics, power management, and more. These peripheral processors
are often single-purpose and are much less powerful than the primary CPU. Built-in
peripherals that don’t implement sufficient security become an easier target for attackers
to exploit, through which they can persistently infect the operating system. Having infected
a peripheral processor firmware, an attacker could target software on the primary CPU or
directly capture sensitive data (For example, an Ethernet device could see the contents of
packets that aren’t encrypted.)
Whenever possible, Apple works to reduce the number of peripheral processors necessary
or works to avoid designs that require firmware. But when separate processors with their
own firmware are required, efforts are taken to ensure an attacker can’t persist on that
processor. This can be by verifying the processor in one of two ways:
• Running the processor so that it downloads verified firmware from the primary CPU on
startup
• Having the peripheral processor implement its own secure boot chain, to verify the
peripheral processor firmware every time the Mac starts up
Apple works with vendors to audit their implementations and enhance their designs to
include desired properties such as:
• Ensuring minimum cryptographic strengths
• Ensuring strong revocation of known bad firmware
• Disabling debug interfaces
• Signing the firmware with cryptographic keys that are stored in Apple-controlled
hardware security modules (HSMs)
In recent years, Apple has worked with some external vendors to adopt the same “Image4”
data structures, verification code, and signing infrastructure used by Apple silicon.
When neither storage-free operation nor storage plus secure boot is an option, the design
mandates that firmware updates be cryptographically signed and verified before the
persistent storage can be updated.
Rosetta 2 on a Mac with Apple silicon

A Mac with Apple silicon is capable of running code compiled for the x86_64 instruction set
using a translation mechanism called Rosetta 2. There are two types of translation offered:
just in time and ahead of time.
Just-in-time translation
In the just-in-time (JIT) translation pipeline, an x86_64 Mach object is identified early
in the image execution path. When these images are encountered, the kernel transfers
control to a special Rosetta translation stub rather than to the dynamic link editor,
dyld(1). The translation stub then translates x86_64 pages during the image’s execution.
This translation takes place entirely within the process. The kernel still verifies the code
hashes of each x86_64 page against the code signature attached to the binary as the page
is faulted in. In the event of a hash mismatch, the kernel enforces the remediation policy
appropriate for that process.

Ahead-of-time translation
In the ahead-of-time (AOT) translation path, x86_64 binaries are read from storage at
times the system deems optimal for responsiveness of that code. The translated artifacts
are written to storage as a special type of Mach object file. That file is similar to an
executable image, but it’s marked to indicate it’s the translated product of another image.
In this model, the AOT artifact derives all of its identity information from the original
x86_64 executable image. To enforce this binding, a privileged userspace entity signs the
translation artifact using a device-specific key that’s managed by the Secure Enclave. This
key is released only to the privileged userspace entity, which is identified as such using
a restricted entitlement. The code directory created for the translation artifact includes
the code directory hash of the original x86_64 executable image. The signature on the
translation artifact itself is known as the supplemental signature.
The AOT pipeline begins similarly to the JIT pipeline, with the kernel transferring control
to the Rosetta runtime rather than to the dynamic link editor, dyld(1). But the Rosetta
runtime then sends an interprocess communication (IPC) query to the Rosetta system
service, which asks whether there’s an AOT translation available for the current executable
image. If found, the Rosetta service provides a handle to that translation, and it’s mapped
into the process and executed. During execution, the kernel enforces the code directory
hashes of the translation artifact which are authenticated by the signature rooted in the
device-specific signing key. The original x86_64 image’s code directory hashes aren’t
involved in this process.
Translated artifacts are stored in a Data Vault which isn’t runtime-accessible by any entity
except for the Rosetta service. The Rosetta service manages access to its cache by
distributing read-only file descriptors to individual translation artifacts; this limits access to
the AOT artifact cache. This service’s interprocess communication and dependent footprint
are kept intentionally very narrow to limit its attack surface.
If the code directory hash of the original x86_64 image doesn’t match with the one
encoded into the AOT translation artifact’s signature, this result is considered the
equivalent of an invalid code signature, and appropriate enforcement action is taken.
If a remote process queries the kernel for the entitlements or other code identity
properties of an AOT-translated executable, the identity properties of the original x86_64
image are returned to it.
Static trust cache content

macOS 11 or later ships with Mach “fat” binaries that contain slices of x86_64 and arm64
machine code. On a Mac with Apple silicon, the user may decide to execute the x86_64
slice of a system binary through the Rosetta pipeline—for example to load a plug-in that
has no native arm64 variant. To support this approach, the static trust cache that ships
with macOS, generally, contains three code directory hashes per Mach object file:
• A code directory hash of the arm64 slice
• A code directory hash of the x86_64 slice
• A code directory hash of the AOT translation of the x86_64 slice
The Rosetta AOT translation procedure is deterministic in that it reproduces identical

output for any given input, irrespective of when the translation was performed or on what
device it was performed.

During the macOS build, every Mach object file is run through the Rosetta AOT translation
pipeline associated with the version of macOS being built, and the resulting code directory
hash is recorded into the trust cache. For efficiency, the actual translated products don’t
ship with the operating system and are reconstituted on demand when the user requests
them.
When an x86_64 image is being executed on a Mac with Apple silicon, if that image’s code
directory hash is in the static trust cache, the resulting AOT artifact’s code directory hash
is also expected to be in the static trust cache. Such products aren’t signed by the device-
specific key, because the signing authority is rooted in the Apple secure boot chain.
Unsigned x86_64 code

A Mac with Apple silicon doesn’t permit native arm64 code to execute unless a valid
signature is attached. This signature can be as simple as an ad hoc code signature
(cf. codesign(1)) that doesn’t bear any actual identity from the secret half of an
asymmetric key pair (it’s simply an unauthenticated measurement of the binary).
For binary compatibility, translated x86_64 code is permitted to execute through Rosetta
with no signature information at all. No specific identity is conveyed to this code through
the device-specific Secure Enclave signing procedure, and it executes with precisely the
same limitations that native unsigned code executing on an Intel-based Mac.
Direct memory access protections for Mac computers

To achieve high throughput on high-speed interfaces like PCIe, FireWire, Thunderbolt, and
USB, computers must support direct memory access (DMA) from peripherals. That is, they
must be able to read and write to RAM without continuous involvement of the CPU. Since
2012, Mac computers have implemented numerous technologies to protect DMA, resulting
in the best and most comprehensive set of DMA protections on any PC.
Direct memory access protections for a Mac with Apple silicon

Apple systems on chip contain an Input/Output Memory Management Unit (IOMMU)
for each DMA agent in the system, including PCIe and Thunderbolt ports. Because
each IOMMU has its own set of address translation tables to translate DMA requests,
peripherals connected by PCIe or Thunderbolt can access only memory that has been
explicitly mapped for their use. Peripherals can’t access memory belonging to other parts
of the system—such as the kernel or firmware—memory assigned to other peripherals. If
an IOMMU detects an attempt by a peripheral to access memory that isn’t mapped for that
peripheral’s use, it triggers a kernel panic.
Direct memory access protections for an Intel-based Mac

On an Intel-based Mac with Intel Virtualization Technology for Directed I/O (VT-d) initialize
the IOMMU, enabling DMA remapping and interrupt remapping very early in the boot
process to mitigate various classes of security vulnerabilities. The Apple IOMMU hardware
begins operation with a default-deny policy, so the instant when the system is powered on,
it automatically begins blocking DMA requests from peripherals. After being initialized by
software, the IOMMUs begin allowing DMA requests from peripherals to memory regions
that have been explicitly mapped for their use.

Note: Interrupt remapping for PCIe isn’t necessary on a Mac with Apple silicon, because
each IOMMU handles MSIs for its own peripherals.
Starting in macOS 11, all Mac computers with an Apple T2 Security Chip run UEFI drivers
that facilitate DMA in a restricted ring 3 environment when these drivers are pairing with
external devices. This property helps mitigate security vulnerabilities that may occur
when a malicious device interacts with a UEFI driver in an unexpected way at boot time. In
particular, it reduces the impact of vulnerabilities in a drivers handling of DMA buffers.
Kernel extensions in macOS

Starting with macOS 11, if third-party kernel extensions (kexts) are enabled, they can’t
be loaded into the kernel on demand. Instead, they’re merged into an Auxiliary Kernel
Collection (AuxKC), which is loaded during the boot process. For a Mac with Apple silicon,
the measurement of the AuxKC is signed into the LocalPolicy (for previous hardware, the
AuxKC resided on the data volume). Rebuilding the AuxKC requires the user’s approval and
restarting of the macOS to load the changes into the kernel, and it requires that the secure
boot be configured to Reduced Security.
Important: Kexts are no longer recommended for macOS. Kexts risk the integrity and
reliability of the operating system, and Apple recommends users select solutions that don’t
require extending the kernel.
Kernel extensions in a Mac with Apple silicon

Kexts must be explicitly enabled for a Mac with Apple silicon by holding the power button
at startup to enter into One True Recovery (1TR) mode, then downgrading to Reduced
Security and checking the box to enable kernel extensions. This action also requires
entering an administrator password to authorize the downgrade. The combination of the
1TR and password requirement makes it difficult for software-only attackers starting
from within macOS to inject kexts into macOS, which they can then exploit to gain kernel
privileges.
After a user authorizes kexts to load, the above User-Approved Kernel Extension Loading
flow is used to authorize the installation of kexts. The authorization used for the above
flow is also used to capture an SHA384 hash of the user-authorized kext list (UAKL) in the
LocalPolicy. The kernel management daemon (kmd) is then responsible for validating only
those kexts found in the UAKL for inclusion into the AuxKC.
• If System Integrity Protection (SIP) is enabled, the signature of each kext is verified
before being included in the AuxKC.
• If SIP is disabled, the kext signature isn’t enforced.
This approach allows Permissive Security flows for developers or users who aren’t part of
the Apple Developer Program to test kexts before they are signed.

After the AuxKC is created, its measurement is sent to the Secure Enclave to be signed
and included in an Image4 data structure that can be evaluated by iBoot at startup. As part
of the AuxKC construction, a kext receipt is also generated. This receipt contains the list
of kexts that were actually included in the AuxKC, because the set could be a subset of
the UAKL if banned kexts were encountered. An SHA384 hash of the AuxKC Image4 data
structure and the kext receipt are included in the LocalPolicy. The AuxKC Image4 hash
is used for extra verification by iBoot at startup to ensure that it isn’t possible to start
up an older Secure Enclave–signed AuxKC Image4 file with a newer LocalPolicy. The kext
receipt is used by subsystems such as Apple Pay to determine whether there are any kexts
currently loaded that could interfere with the trustworthiness of macOS. If there are, then
Apple Pay capabilities may be disabled.
Alternatives to kexts (macOS 10.15 or later)

macOS 10.15 enables developers to extend the capabilities of macOS by installing and
managing system extensions that run in user space rather than at the kernel level. By
running in user space, system extensions increase the stability and security of macOS.
Even though kexts inherently have full access to the entire operating system, extensions
running in user space are granted only the privileges necessary to perform their specified
function.
Developers can use frameworks, including DriverKit, EndpointSecurity, and

NetworkExtension, to write USB and human interface drivers, endpoint security tools (like
data loss prevention or other endpoint agents), and VPN and network tools, all without
needing to write kexts. Third-party security agents should be used only if they take
advantage of these APIs or have a robust road map to transition to them and away from
kernel extensions.
User-Approved Kernel Extension Loading

To improve security, user consent is required to load kernel extensions installed with or
after installing macOS 10.13. This process is known as User-Approved Kernel Extension
Loading. Administrator authorization is required to approve a kernel extension. Kernel
extensions don’t require authorization if they:
• Were installed on a Mac when running macOS 10.12 or earlier
• Are replacing previously approved extensions
• Are allowed to load without user consent by using the spctl command-line tool
available when a Mac was booted from recoveryOS
• Are allowed to load using mobile device management (MDM) configuration
Starting with macOS 10.13.2, users can use MDM to specify a list of kernel extensions
that load without user consent. This option requires a Mac running macOS 10.13.2 that’s
enrolled in MDM—through Apple School Manager, Apple Business Manager, or user-
approved MDM enrollment.

Option ROM security in macOS
Note: Option ROMs aren’t currently supported on a Mac with Apple silicon.
Option ROM security in a Mac with the Apple T2 Security Chip

Both Thunderbolt and PCIe devices can have an “Option ROM (OROM)” physically attached
to the device. (This is typically not a true ROM but is instead a rewritable chip that stores
firmware.) On UEFI-based systems, that firmware is typically a UEFI driver, which is read
in by the UEFI firmware and executed. The executed code is supposed to initialize and
configure the hardware it was retrieved from, so that the hardware can be made usable by
the rest of the firmware. This capability is required so that specialized third-party hardware
can load and operate during the earliest startup phases—for example, to start up from
external RAID arrays.
However, because OROMs are generally rewritable, if an attacker overwrites the OROM of a
legitimate peripheral, the attacker’s code executes early in the boot process and is able to
tamper with the execution environment and violate the integrity of software that’s loaded
later. Likewise, if the attacker introduces their own malicious device to the system, they’re
also able to execute malicious code.
In macOS 10.12.3, the behavior of Mac computers sold after 2011 was changed to not
execute OROMs by default at the time the Mac booted unless a special key combination
was pressed. This key combination protected against malicious OROMs being inadvertently
introduced into the macOS boot sequence. The default behavior of the Firmware Password
Utility was also changed so that when the user set a firmware password, OROMs couldn’t
execute even if the key combination was pressed. This protected against a physically
present attacker intentionally introducing a malicious OROM. For users who still need
to run OROMs while they have a firmware password set, a nondefault option can be
configured using the firmwarepasswd command-line tool in macOS.

OROM sandbox security
In macOS 10.15, UEFI firmware was updated to contain a mechanism for sandboxing
OROMs and for stripping privileges from them. UEFI firmware typically executes all code,
including OROMs, at the maximum CPU privilege level, called ring 0, and has a single
shared virtual memory space for all code and data. Ring 0 is the privilege level where the
macOS kernel runs, whereas the lower privilege level, ring 3, is where apps run. The OROM
sandbox deprivileged OROMs by making use of virtual memory separation like the kernel
does and then making the OROMs run in ring 3.
Option ROM (OROM) sandboxing.
The sandbox further significantly restricts both the interfaces that the OROMs can call
(much like system call filtering in kernels) and the type of device that an OROM can
register as (much like app approval.) The benefit of this design is that malicious OROMs
can no longer directly write anywhere within ring 0 memory. Instead, they are limited to a
very narrow and well-defined sandbox interface. This limited interface significantly reduces
attack surface and forces attackers to first escape the sandbox and escalate privilege.

UEFI firmware security in an Intel-based Mac
Overview
Since 2006, Mac computers with an Intel-based CPU use an Intel firmware based on the
Extensible Firmware Interface (EFI) Development Kit (EDK) version 1 or version 2. EDK2-
based code conforms to the Unified Extensible Firmware Interface (UEFI) specification.
This section refers to the Intel firmware as the UEFI firmware. The UEFI firmware was the
first code to execute on the Intel chip.
For an Intel-based Mac without the Apple T2 Security Chip, the root of trust for the UEFI
firmware is the chip where the firmware is stored. UEFI firmware updates are digitally
signed by Apple and verified by the firmware before updating the storage. To prevent
rollback attacks, updates must always have a version newer than the existing one. However,
an attacker with physical access to the Mac could potentially use hardware to attach to
the firmware storage chip and update the chip to contain malicious content. Likewise, if
vulnerabilities are found in the early boot process of the UEFI firmware (before it write-
restricts the storage chip), this could also lead to persistent infection of the UEFI firmware.
This is a hardware architectural limitation common in most Intel-based PCs and present in
all Intel-based Mac computers without the T2 chip.
To help prevent physical attacks that subvert UEFI firmware, Mac computers were
rearchitected to root the trust in the UEFI firmware in the T2 chip. On these Mac computers,
the root of trust for the UEFI firmware is specifically the T2 firmware, as described in Boot
process for an Intel-based Mac.
Intel Management Engine (ME) subcomponent

One subcomponent stored within the UEFI firmware is the Intel Management Engine
(ME) firmware. The ME—a separate processor and subsystem within Intel chips—is used
primarily for audio and video copyright protection on a Mac that has only Intel-based
graphics. To reduce this subcomponent’s attack surface, an Intel-based Mac runs a custom
ME firmware from which most components have been removed. Because the resulting
Mac ME firmware is smaller than the default minimal build that Intel makes available, many
components that have been the subject of public attacks by security researchers in the
past are no longer present.
System Management Mode (SMM)

Intel processors have a special execution mode that’s distinct from normal operation.
Called System Management Mode (SMM), it was originally introduced to handle time-
sensitive operations such as power management. However, to perform such actions, Mac
computers have historically used a discrete microcontroller called the System Management
Controller (SMC). No longer a separate microcontroller, the SMC has been integrated into
the T2 chip.

System security for watchOS
Apple Watch uses many of the same hardware-based platform security capabilities to
perform secure boot, secure software updates, maintain operating system integrity and
help protect data on the device as well as during communication with its paired iPhone
and with the internet. Supported technologies include those listed in System Security (for
example, KIP and SCIP) as well as Data Protection, keychain, and network technologies.
Secure pairing with iPhone

Apple Watch can be paired with only one iPhone at a time. When Apple Watch is unpaired,
iPhone communicates instructions to erase all content and data from the watch.
Pairing Apple Watch with iPhone is secured using an out-of-band process to exchange
public keys, followed by the Bluetooth Low Energy (BLE) link shared secret. Apple Watch
displays an animated pattern, which is captured by the camera on iPhone. The pattern
contains an encoded secret that’s used for BLE 4.1 out-of-band pairing. Standard BLE
Passkey Entry is used as a fallback pairing method, if necessary.
After the BLE session is established and encrypted using the highest security protocol
available in the Bluetooth Core Specification, iPhone and Apple Watch exchange keys using
either:
• A process adapted from Apple Identity Service (IDS) as described in the iMessage
security overview.
• A key exchange using IKEv2/IPsec. The initial key exchange is authenticated using either
the Bluetooth session key (for pairing scenarios) or the IDS keys (for operating system
update scenarios). Each device generates a random public and private 256-bit Ed25519
key pair, and during the initial key exchange process, the public keys are exchanged.
The mechanism used for key exchange and encryption depends on which operating system
versions are on the iPhone and Apple Watch. iPhone devices running iOS 13 or later when
paired with an Apple Watch running watchOS 6 or later use only IKEv2/IPsec for key
exchange and encryption.
After keys have been exchanged:
• The Bluetooth session key is discarded and all communications between iPhone and
Apple Watch are encrypted using one of the methods listed above—with the encrypted
Bluetooth, Wi-Fi, and cellular links providing a secondary encryption layer.
• (IKEv2/IPsec only) The keys are stored in the system keychain and used for
authenticating future IKEv2/IPsec sessions between the devices. Further
communication between these devices is encrypted and integrity protected using
ChaCha20-Poly1305 (256-bit keys).
The Bluetooth Low Energy device address is rotated at 15-minute intervals to reduce the
risk of local tracking of the device using the broadcast of a persistent identifier.
To support apps that need streaming data, encryption is provided using methods described
in FaceTime security, using either the Apple Identity Service (IDS) provided by the paired
iPhone or a direct internet connection.

Apple Watch implements hardware-encrypted storage and class-based protection of files
and keychain items. Access-controlled keybags for keychain items are also used. Keys
used to communicate between Apple Watch and iPhone are also secured using class-
based protection. For more information, see Keybags for Data Protection.
Secure use of Wi-Fi, cellular, iCloud, and Gmail

When Apple Watch isn’t within Bluetooth range, Wi-Fi or cellular can be used instead.
Apple Watch automatically joins Wi-Fi networks that have been already been joined on the
paired iPhone and whose credentials have synced to the Apple Watch while both devices
were in range. This Auto-Join behavior can then be configured on a per-network basis in
the Wi-Fi section of the Apple Watch Settings app. Wi-Fi networks that have never been
joined before on either device can be manually joined in Wi-Fi section of the Apple Watch
Settings app.
When Apple Watch and iPhone are out of range, Apple Watch connects directly to iCloud
and Gmail servers to fetch Mail, as opposed to syncing Mail data with the paired iPhone
over the internet. For Gmail accounts, the user is required to authenticate to Google in the
Mail section of the Watch app on iPhone. The OAuth token received from Google is sent
over to Apple Watch in encrypted format over Apple Identity Service (IDS) so it can be
used to fetch Mail. This OAuth token is never used for connectivity with the Gmail server
from the paired iPhone.
Locking and unlocking Apple Watch

If wrist detection is enabled, the device locks automatically shortly after it’s removed from
the user’s wrist. If wrist detection is disabled, Control Center provides an option for locking
Apple Watch. When Apple Watch is locked, Apple Pay can be used only by entering the
watch’s passcode. Wrist detection is turned off using the Apple Watch app on iPhone. This
setting can also be enforced using a mobile device management (MDM) solution.
The paired iPhone can also unlock the watch, provided the watch is being worn. This is
accomplished by establishing a connection authenticated by the keys established during
pairing. iPhone sends the key, which the watch uses to unlock its Data Protection keys. The
watch passcode isn’t known to iPhone nor is it transmitted. This feature can be turned off
using the Apple Watch app on iPhone.
Enabling Find My on the paired iPhone also allows the use of Activation Lock on
Apple Watch. Activation Lock makes it harder for anyone to use or sell an Apple Watch
that’s been lost or stolen. Activation Lock requires the user’s Apple ID and password to
unpair, erase, or reactivate an Apple Watch.

Updating system software
Apple Watch can be configured for a system software update the same night. For more
information on how the Apple Watch passcode gets stored and used during the update,
see Keybags.
Auto Unlock with Apple Watch in macOS

Users with Apple Watch can use it to automatically unlock their Mac. Bluetooth Low
Energy (BLE) and peer-to-peer Wi-Fi allow Apple Watch to securely unlock a Mac after
ensuring proximity between the devices. This requires an iCloud account with two-factor
authentication configured.
When enabling an Apple Watch to unlock a Mac, a secure link using Auto Unlock Identities
is established. The Mac creates a random one-time-use unlock secret and transmits it to
the Apple Watch over the link. The secret is stored on Apple Watch and can be accessed
only when Apple Watch is unlocked. The unlock token isn’t the user’s password.
During an unlock operation, the Mac uses BLE to create a connection to the Apple Watch. A
secure link is then established between the two devices using the shared keys used when
it was first enabled. The Mac and Apple Watch then use peer-to-peer Wi-Fi and a secure
key derived from the secure link to determine the distance between the two devices. If the
devices are within range, the secure link is then used to transfer the preshared secret to
unlock the Mac. After successful unlock, the Mac replaces the current unlock secret with
a new one-time use unlock secret and transmits the new unlock secret to the Apple Watch
over the link.
Approve with Apple Watch

When Auto Unlock with Apple Watch is enabled, the Apple Watch can be used in place or
together with Touch ID to approve authorization and authentication prompts from:
• macOS and Apple apps that request authorization
• Third-party apps that request authentication
• Saved Safari passwords

• Secure Notes

Random number generation
Cryptographic pseudorandom number generators (CPRNGs) are an important building
block for secure software. To this end, Apple provides a trusted software CPRNG running
in the iOS, iPadOS, macOS, tvOS, and watchOS kernels. It’s responsible for aggregating
raw entropy from the system and providing secure random numbers to consumers in both
the kernel and user space.
Entropy sources
The kernel CPRNG is seeded from multiple entropy sources during boot and over the
lifetime of the device. These include (contingent on availability):
• The Secure Enclave hardware TRNG
• Timing-based jitter collected during boot
• Entropy collected from hardware interrupts
• A seed file used to persist entropy across boots
• Intel random instructions—for example, RDSEED and RDRAND (only on an Intel-based

Mac)
The kernel CPRNG

The kernel CPRNG is a Fortuna-derived design targeting a 256-bit security level. It
provides high-quality random numbers to user-space consumers using the following APIs:
• The getentropy(2) system call
• The random device (/dev/random)
The kernel CPRNG accepts user-supplied entropy through writes to the random device.
Apple Security Research Device

The Apple Security Research Device is a specially fused iPhone that allows security
researchers to perform research on iOS without having to defeat or disable the platform
security features of iPhone. With this device, a researcher can side-load content that runs
with platform-equivalent permissions and thus perform research on a platform that more
closely models that of production devices.
To ensure that user devices aren’t affected by the security research device execution
policy, the policy changes are implemented in a variant of iBoot and the Boot Kernel
Collection. These fail to boot on user hardware. The research iBoot checks for a new fusing
state and enters a panic loop if it’s being run on nonresearch fused hardware.
The cryptex subsystem allows a researcher to load a personalized trust cache and a disk
image containing corresponding content. A number of defense in-depth measures have
been implemented to ensure that this subsystem doesn’t allow execution on user devices:
• launchd won’t load the cryptexd launchd property list if it’s unable to detect the
research fuse.
• cryptexd aborts if it doesn’t detect the research fuse.

• The entitlement that grants cryptexd the ability to mount a disk image is honored only
by the research kernel cache. The relevant code path isn’t compiled into the release
kernel cache.
• The signing server refuses to personalize a cryptex disk image for a device not on an
explicit allow list.
To respect the privacy of the security researcher, only the measurements (for example,
hashes) of the executables and the security research device identifiers are sent to Apple
during personalization. Apple doesn’t receive the content of the cryptex being loaded onto
the device.
To avoid having a malicious party attempt to masquerade a research device as a user

device to trick a target into using it for everyday usage, the security research device has
the following differences:
• The security research device starts up only while charging. This can be using a lightning
cable or a Qi-compatible charger. If the device isn’t charging during startup, the device
enters Recovery mode. If the user starts charging and restarts the device, it starts up
as normal. As soon as XNU starts, the device doesn’t need to be charging to continue
operation.
• The words Security Research Device are displayed below the Apple logo during iBoot
startup.
• The XNU kernel boots in verbose mode.
• The device is etched on the side with the message “Property of Apple. Confidential and
Proprietary. Call +1 877 595 1125.”
The following are additional measures that are implemented in software that appears after
boot:
• The words Security Research Device are displayed during device setup.
• The words Security Research Device are displayed on the lock screen and in the
Settings app.
The Security Research Device affords researchers the following abilities that a user device
doesn’t:
• Side-load executable code onto the device with arbitrary entitlements at the same
permission level as Apple operating system components.
• Start services at startup.
• Persist content across restarts.

Encryption and Data Protection
Encryption and Data Protection overview

The secure boot chain, system security, and app security capabilities all help to verify that
only trusted code and apps run on a device. Apple devices have additional encryption
features to safeguard user data, even when other parts of the security infrastructure have
been compromised (for example, if a device is lost or is running untrusted code). All of
these features benefit both users and IT administrators, protecting personal and corporate
information and providing methods for instant and complete remote wipe in the case of
device theft or loss.
iOS and iPadOS devices used a file encryption methodology called Data Protection,
whereas the data on an Intel-based Mac is protected with a volume encryption technology
called FileVault. A Mac with Apple silicon uses a hybrid model that supports Data
Protection, with two caveats: The lowest protection level Class (D) isn’t supported, and
the default level (Class C) uses a volume key and acts just like the FileVault on an Intel-
based Mac. In all cases, key management hierarchies are rooted in the dedicated silicon
of the Secure Enclave, and a dedicated AES Engine supports line-speed encryption and
helps ensure that long-lived encryption keys aren’t exposed to the kernel operating system
or CPU (where they might be compromised). (An Intel-based Mac with a T1 or lacking a
Secure Enclave doesn’t use dedicated silicon to protect its FileVault encryption keys.)
Besides using Data Protection and FileVault to prevent unauthorized access to data, Apple
operating system kernels enforce protection and security. The kernel uses access controls
to sandbox apps (which restricts what data an app can access) and a mechanism called a
Data Vault (which rather than restricting the calls an app can make, restricts access to the
data of an app from all other requesting apps).
Passcodes and passwords

Passcodes in devices that support Data Protection
By setting up a device passcode or password, the user automatically enables Data
Protection. iOS and iPadOS support six-digit, four-digit, and arbitrary-length alphanumeric
passcodes. Besides unlocking the device, a passcode or password provides entropy for
certain encryption keys. This means an attacker in possession of a device can’t get access
to data in specific protection classes without the passcode.

The passcode or password is entangled with the device’s UID, so brute-force attempts
must be performed on the device under attack. A large iteration count is used to
make each attempt slower. The iteration count is calibrated so that one attempt takes
approximately 80 milliseconds. In fact, it would take more than five and one-half years to
try all combinations of a six-character alphanumeric passcode with lowercase letters and
numbers.
The stronger the user passcode is, the stronger the encryption key becomes. And by
using Touch ID and Face ID, the user can establish a much stronger passcode than would
otherwise be practical. The stronger passcode increases the effective amount of entropy
protecting the encryption keys used for Data Protection, without adversely affecting the
user experience of unlocking a device multiple times throughout the day.
To further discourage brute-force passcode attacks, there are escalating time delays after
the entry of an invalid passcode at the Lock Screen.
Delays between passcode attempts

Attempts Delay enforced
1–4 None
5 1 minute
6 5 minutes
7–8 15 minutes
9 1 hour
If the Erase Data option is turned on (in Settings > Touch ID & Passcode), after 10
consecutive incorrect attempts to enter the passcode, all content and settings are
removed from storage. Consecutive attempts of the same incorrect passcode don’t count
toward the limit. This setting is also available as an administrative policy through a mobile
device management (MDM) solution that supports this feature and through Microsoft
Exchange ActiveSync, and can be set to a lower threshold.
On devices with Secure Enclave, the delays are enforced by the Secure Enclave. If the
device is restarted during a timed delay, the delay is still enforced, with the timer starting
over for the current period.
Specifying longer passcodes

If a long password that contains only numbers is entered, a numeric keypad is displayed at
the Lock Screen instead of the full keyboard. A longer numeric passcode may be easier to
enter than a shorter alphanumeric passcode, while providing similar security.
Users can specify a longer alphanumeric passcode by selecting Custom Alphanumeric

Code in the Passcode Options in Settings > Touch ID & Passcode or Face ID & Passcode.

Delays between passwords attempts in macOS
To prevent brute-force attacks, when Mac starts up, no more than 30 password attempts
are allowed at the Login Window or using Target Disk Mode, and escalating time delays are
imposed after incorrect attempts. The delays are enforced by the Secure Enclave. If Mac
is restarted during a timed delay, the delay is still enforced, with the timer starting over for
the current period.
To prevent malware from causing permanent data loss by trying to attack the user’s
password, these limits aren’t enforced after the user has successfully logged into the
Mac, but is reimposed after reboot. If the 30 attempts are exhausted, 10 more attempts
are available after booting into recoveryOS. And if those are also exhausted, then 60
additional attempts are available for each FileVault recovery mechanism (iCloud recovery,
FileVault recovery key, and institutional key), for a maximum of 180 additional attempts.
Once those additional attempts are exhausted, the Secure Enclave no longer processes any
requests to decrypt the volume or verify the password, and the data on the drive becomes
unrecoverable.
To protect data in an enterprise setting, IT should define and enforce FileVault

configuration policies using an MDM solution. Organizations have several options for
managing encrypted volumes, including institutional recovery keys, personal recovery keys
(that can optionally be stored with MDM for escrow), or a combination of both. Key rotation
can also be set as a policy in MDM.
Delays between password attempts on a Mac with Apple

silicon and those with the T2 chip
Attempts Delay enforced
5 1 minute
6 5 minutes
7 15 minutes
8 15 minutes
9 1 hour
10 Disabled
In a Mac with the Apple T2 Security Chip, the password serves a similar function except
that the key generated is used for FileVault encryption rather than Data Protection. macOS
also offers additional password recovery options:
• iCloud recovery
• FileVault recovery
• FileVault institutional key

Data Protection
Data Protection overview
Apple uses a technology called Data Protection to protect data stored in flash storage
on the devices that feature an Apple SoC such as iPhone, iPad, Apple Watch, Apple TV,
and a Mac with Apple silicon. Data Protection allows the device to respond to common
events such as incoming phone calls but also enables a high level of encryption for user
data. Certain system apps (such as Messages, Mail, Calendar, Contacts, Photos) and
Health data values use Data Protection by default. Third-party apps receive this protection
automatically.
Implementation
Data Protection is implemented by constructing and managing a hierarchy of keys and
builds on the hardware encryption technologies built into Apple devices. Data Protection is
controlled on a per-file basis by assigning each file to a class; accessibility is determined
according to whether the class keys have been unlocked. APFS (Apple File System) allows
the file system to further subdivide the keys into a per-extent basis (where portions of a
file can have different keys).
Every time a file on the data volume is created, Data Protection creates a new 256-bit key
(the per-file key) and gives it to the hardware AES Engine, which uses the key to encrypt
the file as it is written to flash storage. On A14 and M1 devices, the encryption uses AES-
256 in XTS mode where the 256-bit per-file-key goes through a Key Derivation Function
(NIST Special Publication 800-108) to derive a 256-bit tweak and a 256-bit cipher key. The
hardware generations of A9 through A13, S5, and S6 use AES-128 in XTS mode where the
256-bit per file key was split to provide a 128-bit tweak and a 128-bit cipher key.
On a Mac with Apple silicon, Data Protection defaults to Class C (see Data Protection
classes), but utilizes a volume key rather than a per-extent or per-file key—effectively
re-creating the security model of FileVault for user data. Users must still opt-in to FileVault
in order to receive the full protection of entangling the encryption key hierarchy with their
password. Developers can also opt in to a higher protection class that uses a per-file or
per-extent key.
Data Protection in Apple devices

On Apple devices with Data Protection, each file is protected with a unique per-file (or per-
extent) key. The key, wrapped using the NIST AED key wrap algorithm, is further wrapped
with one of several class keys, depending on how the file is meant to be accessed. The
wrapped per-file key is then stored in the file’s metadata.
Devices with APFS format may support cloning of files (zero-cost copies using copy-
on-write technology). If a file is cloned, each half of the clone gets a new key to accept
incoming writes so that new data is written to the media with a new key. Over time, the file
may become composed of various extents (or fragments), each mapping to different keys.
However, all of the extents that comprise a file are guarded by the same class key.

When a file is opened, its metadata is decrypted with the file system key, revealing the
wrapped per-file key and a notation on which class protects it. The per-file (or per-extent)
key is unwrapped with the class key and then supplied to the hardware AES Engine, which
decrypts the file as it’s read from flash storage. All wrapped file key handling occurs in
the Secure Enclave; the file key is never directly exposed to the Application Processor. At
startup, the Secure Enclave negotiates an ephemeral key with the AES Engine. When the
Secure Enclave unwraps a file’s keys, they’re rewrapped with the ephemeral key and sent
back to the Application Processor.
The metadata of all files in the data volume file system are encrypted with a random
volume key, which is created when the operating system is first installed or when the
device is wiped by a user. This key is encrypted and wrapped by a key wrapping key that
is known only to the Secure Enclave for long-term storage. The key wrapping key changes
every time a user erases their device. On A9 (and newer) SoCs, Secure Enclave relies upon
entropy, backed by anti-replay systems, to achieve effaceability and to protect its key
wrapping key, among other assets. For more information, see Secure nonvolatile storage.
Just like per-file or per-extent keys, the metadata key of the data volume is never directly
exposed to the Application Processor; the Secure Enclave provides an ephemeral, per-
boot version instead. When stored, the encrypted file system key is additionally wrapped
by an “effaceable key” stored in Effaceable Storage or using a media key-wrapping key,
protected by Secure Enclave anti-replay mechanism. This key doesn’t provide additional
confidentiality of data. Instead, it’s designed to be quickly erased on demand (by the
user with the “Erase All Content and Settings” option, or by a user or administrator
issuing a remote wipe command from a mobile device management (MDM) solution,
Microsoft Exchange ActiveSync, or iCloud). Erasing the key in this manner renders all files
cryptographically inaccessible.
The contents of a file may be encrypted with one or more per-file (or per-extent) keys that
are wrapped with a class key and stored in a file’s metadata, which in turn is encrypted
with the file system key. The class key is protected with the hardware UID and, for some
classes, the user’s passcode. This hierarchy provides both flexibility and performance. For
example, changing a file’s class only requires rewrapping its per-file key, and a change of
passcode just rewraps the class key.

Data Protection classes
When a new file is created on devices supporting Data Protection, it’s assigned a class by
the app that creates it. Each class uses different policies to determine when the data is
accessible. The basic classes and policies are described in the following sections. Apple
silicon-based Macs do not support Class D: No Protection, and a security boundary is
established around logging in and out (not locking or unlocking as on iPhone, iPad, and
iPod touch).
Class Protection type
Class A: Complete Protection (NSFileProtectionComplete)
Class B: Protected Unless Open (NSFileProtectionCompleteU nlessO pen)
Class C: Protected Until First User Authentication (NSFileProtectionCompleteU ntilFirstU ser

Authentication)
Note: macO S uses a volume key to recreate FileVault
protection characteristics.
Class D: No Protection (NSFileProtectionNone)
Note: Not supported on macO S.
Complete Protection
(NSFileProtectionComplete): The class key is protected with a key derived from the
user passcode or password and the device UID. Shortly after the user locks a device
(10 seconds, if the Require Password setting is Immediately), the decrypted class key is
discarded, rendering all data in this class inaccessible until the user enters the passcode
again or unlocks (logs in to) the device using Touch ID or Face ID.
In macOS, shortly after the last user is logged out, the decrypted class key is discarded,
rendering all data in this class inaccessible until a user enters the passcode again or logs
into the the device using Touch ID.
Protected Unless Open

(NSFileProtectionCompleteUnlessOpen): Some files may need to be written while the
device is locked or the user is logged out. A good example of this is a mail attachment
downloading in the background. This behavior is achieved by using asymmetric elliptic
curve cryptography (ECDH over Curve25519). The usual per-file key is protected by a key
derived using One-Pass Diffie-Hellman Key Agreement as described in NIST SP 800-56A.
The ephemeral public key for the Agreement is stored alongside the wrapped per-file key.
The KDF is Concatenation Key Derivation Function (Approved Alternative 1) as described
in 5.8.1 of NIST SP 800-56A. AlgorithmID is omitted. PartyUInfo and PartyVInfo are the
ephemeral and static public keys, respectively. SHA256 is used as the hashing function. As
soon as the file is closed, the per-file key is wiped from memory. To open the file again, the
shared secret is re-created using the Protected Unless Open class’s private key and the
file’s ephemeral public key, which are used to unwrap the per-file key that is then used to
decrypt the file.
In macOS, the private part of NSFileProtectionCompleteUnlessOpen is accessible as long

as any users on the system are logged in or are authenticated.

Protected Until First User Authentication
(NSFileProtectionCompleteUntilFirstUserAuthentication): This class behaves in the same
way as Complete Protection, except that the decrypted class key isn’t removed from
memory when the device is locked or the user logged out. The protection in this class
has similar properties to desktop full-volume encryption, and protects data from attacks
that involve a reboot. This is the default class for all third-party app data not otherwise
assigned to a Data Protection class.
In macOS, this class utilizes a volume key which is accessible as long as the volume is
mounted, and acts just like FileVault.
No Protection
(NSFileProtectionNone): This class key is protected only with the UID, and is kept in
Effaceable Storage. Since all the keys needed to decrypt files in this class are stored
on the device, the encryption only affords the benefit of fast remote wipe. If a file isn’t
assigned a Data Protection class, it is still stored in encrypted form (as is all data on an iOS
and iPadOS device).
This isn’t supported in macOS.
Note: In macOS, for volumes that don’t correspond to a booted operating system, all data
protection classes are accessible as long as the volume is mounted. The default data
protection class is NSFileProtectionCompleteUntilFirstUserAuthentication. Per-extent key
functionality is available to both Rosetta 2 and native apps.
Keybags for Data Protection

The keys for both file and keychain Data Protection classes are collected and managed in
keybags on iOS, iPadOS, watchOS, and tvOS. These operating systems use the following
keybags: user, device, backup, escrow, and iCloud Backup.
User keybag
The user keybag is where the wrapped class keys used in normal operation of the device
are stored. For example, when a passcode is entered, NSFileProtectionComplete is loaded
from the user keybag and unwrapped. It is a binary property list (.plist) file stored in the No
Protection class.
For devices with SoCs earlier than the A9, the .plist file contents are encrypted with a
key held in Effaceable Storage. To give forward security to keybags, this key is wiped and
regenerated each time a user changes their passcode.
For devices with the A9 or later SoCs, the .plist file contains a key that indicates that the
keybag is stored in a locker protected by the Secure Enclave–controlled anti-replay nonce.
The Secure Enclave manages the user keybag and can be queried regarding a device’s lock
state. It reports that the device is unlocked only if all the class keys in the user keybag are
accessible and have been unwrapped successfully.

Device keybag
The device keybag is used to store the wrapped class keys used for operations involving
device-specific data. iOS and iPadOS devices configured for shared use sometimes
need access to credentials before any user has logged in; therefore, a keybag that isn’t
protected by the user’s passcode is required.
iOS and iPadOS don’t support cryptographic separation of per-user file system content,
which means the system uses class keys from the device keybag to wrap per-file keys.
The keychain, however, uses class keys from the user keybag to protect items in the
user keychain. In iOS and iPadOS devices configured for use by a single user (the default
configuration), the device keybag and the user keybag are one and the same, and are
protected by the user’s passcode.
Backup keybag
The backup keybag is created when an encrypted backup is made by the Finder
(macOS 10.15 or later) or iTunes (in macOS 10.14 or earlier) and stored on the computer
to which the device is backed up. A new keybag is created with a new set of keys, and the
backed-up data is reencrypted to these new keys. As explained previously, nonmigratory
keychain items remain wrapped with the UID-derived key, allowing them to be restored
to the device they were originally backed up from but rendering them inaccessible on a
different device.
The keybag—protected with the password set—is run through 10 million iterations of
the key derivation function PBKDF2. Despite this large iteration count, there’s no tie to a
specific device, and therefore a brute-force attack parallelized across many computers
could theoretically be attempted on the backup keybag. This threat can be mitigated with a
sufficiently strong password.
If a user chooses not to encrypt the backup, the files aren’t encrypted regardless of their
Data Protection class but the keychain remains protected with a UID-derived key. This is
why keychain items migrate to a new device only if a backup password is set.
Escrow keybag
The escrow keybag is used for syncing with iTunes (in macOS 10.14 or earlier) or the Finder
(macOS 10.15 or later) through USB and restored to factory default settings. syncing and
mobile device management (MDM). This keybag allows iTunes or the Finder to back up
and sync without requiring the user to enter a passcode, and it allows an MDM solution
to remotely clear a user’s passcode. It is stored on the computer that’s used to sync with
iTunes or the Finder, or on the MDM solution that remotely manages the device.
The escrow keybag improves the user experience during device syncing, which potentially
requires access to all classes of data. When a passcode-locked device is first connected
to the Finder or iTunes, the user is prompted to enter a passcode. The device then creates
an escrow keybag containing the same class keys used on the device, protected by a newly
generated key. The escrow keybag and the key protecting it are split between the device
and the host or server, with the data stored on the device in the Protected Until First User
Authentication class. This is why the device passcode must be entered before the user
backs up with the Finder or iTunes for the first time after a reboot.

In the case of an over-the-air (OTA) software update, the user is prompted for their
passcode when initiating the update. This is used to securely create a one-time unlock
token, which unlocks the user keybag after the update. This token can’t be generated
without entering the user’s passcode, and any previously generated token is invalidated if
the user’s passcode changed.
One-time unlock tokens are either for attended or unattended installation of a software
update. They’re encrypted with a key derived from the current value of a monotonic
counter in the Secure Enclave, the UUID of the keybag, and the Secure Enclave UID.
On A9 (and later) SoCs, one-time Unlock token no longer relies on counters or Effaceable
Storage. Instead, it’s protected by Secure Enclave controlled anti-replay nonce.
The one-time unlock token for attended software updates expires after 20 minutes. In
iOS 13 and iPadOS 13.1 or later, the token is stored in a locker protected by the Secure
Enclave. Prior to iOS 13, this token was exported from the Secure Enclave and written
to Effaceable Storage or was protected by the Secure Enclave anti-replay mechanism. A
policy timer incremented the counter if the device hadn’t rebooted within 20 minutes.
Unattended software updates occur when the system detects an update is available and
when one of the following is true:
• Automatic updates are configured in iOS 12 or later.
• The user chooses Install Later when notified of the update.
After the user enters their passcode, a one-time unlock token is generated and can remain
valid in Secure Enclave for up to 8 hours. If the update hasn’t yet occurred, this one-time
unlock token is destroyed on every lock and re-created on every subsequent unlock. Each
unlock restarts the 8 hour window. After 8 hours a policy timer invalidates the one-time
unlock token.
iCloud Backup keybag

The iCloud Backup keybag is similar to the backup keybag. All the class keys in this keybag
are asymmetric (using Curve25519, like the Protected Unless Open Data Protection class).
An asymmetric keybag is also used for the backup in the keychain recovery aspect of
iCloud Keychain.

Protecting keys in alternate boot modes
Data Protection is designed to provide access to user data only after successful
authentication, and only to the authorized user. Data protection classes are designed to
support a variety of use cases, such as the ability to read and write some data even when
a device is locked (but after first unlock). Additional steps are taken to protect access to
user data during alternate boot modes such as those used for Device Firmware Update
(DFU) mode, Recovery mode, Apple Diagnostics, or even during software update. These
capabilities are based on a combination of hardware and software features, and have been
expanded as Apple-designed silicon has evolved.
Recovery: All
Data Protection
Classes
protected
Alternate boots
of DFU mode,
Recovery,
and software
updates: Class
A, B, and C data
protected
The Secure Enclave AES Engine is equipped with lockable software seed bits. When keys
are created from the UID, these seed bits are included in the key derivation function to
create additional key hierarchies. How the seed bit is used varies according to the system
on chip:
• Starting with the Apple A10 and S3 SoCs, a seed bit is dedicated to distinguish keys
protected by the user’s passcode. The seed bit is set for keys that require the user’s
passcode (including Data Protection Class A, Class B, and Class C keys), and cleared
for keys that don’t require the user’s passcode (including the file system metadata key
and Class D keys).
• In iOS 13 or later and iPadOS 13.1 or later on devices with an A10 or later, all user data
is rendered cryptographically inaccessible when devices are booted into Diagnostics
mode. This is achieved by introducing an additional seed bit whose setting governs
the ability to access the media key, which itself is needed to access the metadata (and
therefore contents of) all files on the data volume encrypted with Data Protection. This
protection encompasses files protected in all classes (A, B, C, and D), not just those
that required the user’s passcode.
• On A12 SoCs, the Secure Enclave Boot ROM locks the passcode seed bit if the
Application Processor has entered Device Firmware Upgrade (DFU) mode or Recovery
mode. When the passcode seed bit is locked, no operation to change it is allowed,
preventing access to data protected with the user’s passcode.
Restoring a device after it enters DFU mode returns it to a known good state with the
certainty that only unmodified Apple-signed code is present. DFU mode can be entered
manually.

See the following Apple Support articles on how to place a device in DFU mode:
Device Article
iP hone, iPad, iPod touch If you forgot the passcode on your iPhone, or your
iPhone is disabled
Apple TV Restore your Apple TV
A Mac with Apple silicon Revive or restore a Mac with Apple silicon
Protecting user data in the face of attack

Attackers attempting to extract user data often try a number of techniques: extracting
the encrypted data to another medium for brute-force attack, manipulating the operating
system version, or otherwise changing or weakening the security policy of the device to
facilitate attack. Attacking data on a device often requires communicating with the device
using physical interfaces like Lightning or USB. Apple devices include features to help
prevent such attacks.
Apple devices support a technology called Sealed Key Protection (SKP) that works to
ensure cryptographic material is rendered unavailable off device, or if manipulations
are made to operating system version or security settings without appropriate user
authorization. This feature is not provided by the Secure Enclave, instead it is supported
by hardware registers that exist at a lower layer in order to provide an additional layer of
protection to the keys necessary to decrypt user data independent of the Secure Enclave.
Note: SKP is available only on devices with an Apple-designed SoC.
Sealed Key
Protection
iPhone and iPad can also be configured to only activate data connections in conditions
more likely to indicate the device is still under the physical control of the authorized owner.

Sealed Key Protection (SKP)
On Apple devices that support Data Protection, the key encryption key (KEK) is protected
(or sealed) with measurements of the software on the system, as well as being tied to the
UID available only from the Secure Enclave. On a Mac with Apple silicon, the protection of
the KEK is further strengthened by incorporating information about security policy on the
system, because macOS supports critical security policy changes (for example, disabling
secure boot or SIP) that are unsupported on other platforms. On a Mac with Apple silicon,
this protection encompass FileVault keys, because FileVault is implemented using Data
Protection (Class C).
The Sealed Key Protection process.
The Secure Enclave Boot Monitor captures the measurement of the Secure Enclave OS
that is loaded. When the Application Processor Boot ROM measures the Image4 manifest
attached to LLB, that manifest contains a measurement of all other system-paired firmware
that is loaded as well. The LocalPolicy contains the core security configurations for the
macOS which are loaded. The LocalPolicy also contains the nsih field which is a hash
of the macOS Image4 manifest. The macOS Image4 manifest contains measurements of
all of the macOS-paired firmware and core macOS boot objects such as the Boot Kernel
Collection or signed system volume (SSV) root hash.
If an attacker is able to unexpectedly change any of the above measured firmware,

software, or security configuration components, it modifies the measurements stored in
the hardware registers. The modification of the measurements cause the crypto-hardware-
derived system measurement root key (SMRK) to derive to a different value, effectively
breaking the seal on the key hierarchy. That causes the system measurement device
key (SMDK) to be inaccessible, which in turn causes the KEK, and thus the data, to be
inaccessible.

However, the system must accommodate legitimate software updates that change the
firmware measurements and the nsih field in the LocalPolicy to point at new macOS
measurements. In other systems that attempt to incorporate firmware measurements but
that don’t have a known good source of truth, the user is required to disable the security,
update firmware, and then reenable so that a new measurement baseline can be captured.
This significantly increases the risk that the attacker could tamper with firmware during
a software update. The system is helped by the fact that the Image4 manifest contains
all the measurements needed. The hardware that decrypts the SMDK with the SMRK
when the measurements match during a normal boot can also encrypt the SMDK to to a
proposed future SMRK. By specifying the measurements that are expected after a software
update, the hardware can encrypt an SMDK, which is accessible in a current operating
system, so that it remains accessible in a future operating system. Similarly, when a
customer legitimately changes their security settings in the LocalPolicy, the SMDK must be
encrypted to the future SMRK based on the measurement for the LocalPolicy, which LLB
computes on the next restart.
Activating data connections securely in iOS and iPadOS

On iOS or iPadOS devices, if no data connection has been established recently, users
must use Touch ID, Face ID, or a passcode to activate data connections through a
Lightning, USB, or Smart Connector interface. This limits the attack surface against
physically connected devices such as malicious chargers while still enabling usage of other
accessories within reasonable time constraints. If more than an hour has passed since
the iOS or iPadOS device has locked or since an accessory’s data connection has been
terminated, the device won’t allow any new data connections to be established until the
device is unlocked. During this hour period, only data connections from accessories that
have been previously connected to the device while in an unlocked state will be allowed.
These accessories are remembered for 30 days after the last time they were connected.
Attempts by an unknown accessory to open a data connection during this period will
disable all accessory data connections over Lighting, USB, and Smart Connector until the
device is unlocked again. This hour period:
• Ensures that frequent users of connections to a Mac or PC, to accessories, or wired to

CarPlay won’t need to input their passcodes every time they attach their device
• Is necessary because the accessory ecosystem doesn’t provide a cryptographically

reliable way to identify accessories before establishing a data connection
In addition, if it’s been more than 3 days since a data connection has been established with
an accessory, the device will disallow new data connections immediately after it locks. This
is to increase protection for users that don’t often make use of such accessories. Data
connections over Lightning, USB, and Smart Connector are also disabled whenever the
device is in a state where it requires a passcode to reenable biometric authentication.
The user can choose to reenable always-on data connections in Settings (setting up some
assistive devices does this automatically).

Role of Apple File System
Apple File System (APFS) is a proprietary file system that was designed with encryption
in mind. APFS works across all Apple’s platforms—for iPhone, iPad, iPod touch, Mac,
Apple TV, and Apple Watch. Optimized for Flash/SSD storage, it features strong encryption,
copy-on-write metadata, space sharing, cloning for files and directories, snapshots, fast
directory sizing, atomic safe-save primitives, and improved file system fundamentals,
as well as a unique copy-on-write design that uses I/O coalescing to deliver maximum
performance while ensuring data reliability.
Space sharing
APFS allocates storage space on demand. When a single APFS container has multiple
volumes, the container’s free space is shared and can be allocated to any of the individual
volumes as needed. Each volume uses only part of the overall container, so the available
space is the total size of the container, minus the space used in all volumes in the
container.
Multiple volumes
In macOS 10.15 or later, an APFS container used to start up the Mac must contain at least
five volumes, the first three of which are hidden from the user:
• Preboot volume: Contains data needed for booting each system volume in the container
• VM volume: Used by macOS for swap file storage
• Recovery volume: Contains recoveryOS
• System volume: Contains the following:
• All the necessary files to start up the Mac
• All apps installed natively by macOS (apps that used to reside in the /Applications
folder now reside in /System/Applications)
Note: By default, no process can write to the System volume, even Apple system
processes.
• Data volume: Contains data that is subject to change, such as:

• Any data inside the user’s folder, including photos, music, videos, and documents
• Apps the user installed, including AppleScript and Automator applications
• Custom frameworks and daemons installed by the user, organization, or third-party

apps
• Other locations owned and writable by the user, as /Applications, /Library, /Users, /
Volumes, /usr/local, /private, /var, and /tmp
A data volume is created for each additional system volume. The preboot, VM, and
recovery volumes are all shared and not duplicated.
In macOS 11, the system volume is captured in a snapshot. The operating system boots
from a snapshot of the system volume, not just from a read-only mount of the mutable
system volume.

In iOS and iPadOS, storage is divided into at least two APFS volumes:
• System volume
• Data volume
Keychain data protection

Many apps need to handle passwords and other short but sensitive bits of data, such
as keys and login tokens. The keychain provides a secure way to store these items. The
various Apple operating systems use differing mechanisms to enforce the guarantees
associated with the different keychain protection classes. In macOS (including a Mac with
Apple silicon), Data Protection is not used directly to enforce these guarantees.
Overview
Keychain items are encrypted using two different AES-256-GCM keys: a table key
(metadata) and a per-row key (secret key). Keychain metadata (all attributes other than
kSecValue) is encrypted with the metadata key to speed searches, and the secret value
(kSecValueData) is encrypted with the secret key. The metadata key is protected by the
Secure Enclave but is cached in the Application Processor to allow fast queries of the
keychain. The secret key always requires a round trip through the Secure Enclave.
The keychain is implemented as a SQLite database, stored on the file system. There is only
one database, and the securityd daemon determines which keychain items each process
or app can access. Keychain Access APIs result in calls to the daemon, which queries
the app’s “Keychain-access-groups,” “application-identifier,” and “application-group”
entitlements. Rather than limiting access to a single process, access groups allow keychain
items to be shared between apps.
Keychain items can be shared only between apps from the same developer. To share
keychain items, third-party apps to use access groups with a prefix allocated to them
through the Apple Developer Program in their application groups. The prefix requirement
and application group uniqueness are enforced through code signing, provisioning profiles,
and the Apple Developer Program.
Keychain data is protected using a class structure similar to the one used in file Data
Protection. These classes have behaviors equivalent to file Data Protection classes but use
distinct keys and functions.

Availability File data protection Keychain data protection
When unlocked NSFileProtectionComplete kSecAttrAccessibleWhenU nlocked
While locked NSFileProtectionCompleteU nless N/A
Open
After first unlock NSFileProtectionCompleteU ntil kSecAttrAccessibleAfterF irstU nlock
FirstUserAuthentication
Always NSFileProtectionNone kSecAttrAccessibleAlways
Passcode enabled N/A kSecAttrAccessibleWhenPasscode

SetT hisDeviceO nly
Apps that use background refresh services can use kSecAttrAccessibleAfterFirstUnlock for
keychain items that need to be accessed during background updates.
The class kSecAttrAccessibleWhenPasscodeSetThisDeviceOnly behaves the same

as kSecAttrAccessibleWhenUnlocked; however, it’s available only when the device is
configured with a passcode. This class exists only in the system keybag; they:
• Don’t sync to iCloud Keychain
• Aren’t backed up
• Aren’t included in escrow keybags
If the passcode is removed or reset, the items are rendered useless by discarding the class
keys.
Other keychain classes have a “This device only” counterpart, which is always protected
with the UID when being copied from the device during a backup, rendering it useless
if restored to a different device. Apple has carefully balanced security and usability by
choosing keychain classes that depend on the type of information being secured and when
it’s needed by iOS and iPadOS. For example, a VPN certificate must always be available so
the device keeps a continuous connection, but it’s classified as “nonmigratory,” so it can’t
be moved to another device.

Keychain data class protections
The class protections listed below are enforced for keychain items.
Item Accessible
Wi-Fi passwords After first unlock
Mail accounts After first unlock
Microsoft Exchange ActiveS ync accounts After first unlock
VPN passwords After first unlock
LDAP, CalDAV, CardDAV After first unlock
Social network account tokens After first unlock
Handoff advertisement encryption keys After first unlock
iC loud token After first unlock
Home sharing password When unlocked
Safari passwords When unlocked
Safari bookmarks When unlocked
Finder/iTunes backup When unlocked, nonmigratory
VPN certificates Always, nonmigratory
Bluetooth® keys Always, nonmigratory
Apple Push Notification service (APNs) token Always, nonmigratory
iC loud certificates and private key Always, nonmigratory
iM essage keys Always, nonmigratory
Certificates and private keys installed by a Always, nonmigratory

configuration profile
SIM PIN Always, nonmigratory
Find My token Always
Voicemail Always
Keychain access control

Keychains can use access control lists (ACLs) to set policies for accessibility and
authentication requirements. Items can establish conditions that require user presence
by specifying that they can’t be accessed unless authenticated using Touch ID, Face ID,
or by entering the device’s passcode or password. Access to items can also be limited by
specifying that Touch ID or Face ID enrollment hasn’t changed since the item was added.
This limitation helps prevent an attacker from adding their own fingerprint to access a
keychain item. ACLs are evaluated inside the Secure Enclave and are released to the kernel
only if their specified constraints are met.

Keychain architecture in macOS
macOS also provides access to the keychain to conveniently and securely stores user
names and passwords, digital identities, encryption keys, and secure notes. It can be
accessed by opening the Keychain Access app in /Applications/Utilities/. Using a keychain
eliminates the requirement to enter—or even remember—the credentials for each resource.
An initial default keychain is created for each Mac user, though users can create other
keychains for specific purposes.
In addition to relying on user keychains, macOS relies on a number of system-level

keychains that maintain authentication assets that aren’t user specific, such as network
credentials and public key infrastructure (PKI) identities. One of these keychains, System
Roots, is immutable and stores internet PKI root certificate authority (CA) certificates to
facilitate common tasks like online banking and e-commerce. The user can similarly deploy
internally provisioned CA certificates to managed Mac computers to help validate internal
sites and services.
FileVault
Volume encryption with FileVault in macOS
Mac computers offer FileVault, a built-in encryption capability, to secure all data at rest.
FileVault uses the AES-XTS data encryption algorithm to protect full volumes on internal
and removable storage devices.
FileVault on a Mac with Apple silicon is implemented using Data Protection Class C with a
volume key. On a Mac with the Apple T2 Security Chip as well as a Mac with Apple silicon,
encrypted internal storage devices directly connected to the Secure Enclave leverage
its hardware security capabilities as well as that of the AES engine. After a user turns on
FileVault on a Mac, their credentials are required during the boot process.
Internal storage with FileVault turned on

Without valid login credentials or a cryptographic recovery key, the internal APFS volumes
remain encrypted and are protected from unauthorized access even if the physical storage
device is removed and connected to another computer. In macOS 10.15, this includes
both the system volume and the data volume. Starting in macOS 11, the system volume
is protected by the signed system volume (SSV) feature, but the data volume remains
protected by encryption. Internal volume encryption on a Mac with Apple silicon as well as
those with the T2 chip is implemented by constructing and managing a hierarchy of keys,
and builds on the hardware encryption technologies built into the chip. This hierarchy of
keys is designed to simultaneously achieve four goals:
• Require the user’s password for decryption
• Protect the system from a brute-force attack directly against storage media removed
from Mac
• Provide a swift and secure method for wiping content via deletion of necessary
cryptographic material

• Enable users to change their password (and in turn the cryptographic keys used to
protect their files) without requiring reencryption of the entire volume
Internal volume encryption when FileVault is turned on in macOS.
On a Mac with Apple silicon and those with the T2 chip, all FileVault key handling occurs
in the Secure Enclave; encryption keys are never directly exposed to the Intel CPU. All
APFS volumes are created with a volume encryption key by default. Volume and metadata
contents are encrypted with this volume encryption key, which is wrapped with the class
key. The class key is protected by a combination of the user’s password and the hardware
UID when FileVault is turned on.
Internal storage with FileVault turned off

If FileVault isn’t turned on in a Mac with Apple silicon or a Mac with the T2 chip during the
initial Setup Assistant process, the volume is still encrypted but the volume encryption key
is protected only by the hardware UID in the Secure Enclave.
Internal volume encryption when FileVault is turned off in macOS.
If FileVault is turned on later—a process that is immediate since the data was already
encrypted—an anti-replay mechanism prevents the old key (based on hardware UID only)
from being used to decrypt the volume. The volume is then protected by a combination of
the user password with the hardware UID as previously described.

Deleting FileVault volumes
When deleting a volume, its volume encryption key is securely deleted by the Secure
Enclave. This prevents future access with this key even by the Secure Enclave. In addition,
all volume encryption keys are wrapped with a media key. The media key doesn’t provide
additional confidentiality of data, but instead is designed to enable swift and secure
deletion of data because without it, decryption is impossible.
On a Mac with Apple silicon and those with the T2 chip, the media key is guaranteed to
be erased by the Secure Enclave supported technology—for example by remote MDM
commands. Erasing the media key in this manner renders the volume cryptographically
inaccessible.
Removable storage devices

Encryption of removable storage devices doesn’t utilize the security capabilities of the
Secure Enclave, and its encryption is performed in the same manner as an Intel-based Mac
without the T2 chip.
Managing FileVault in macOS
Using Secure Token

Apple File System (APFS) in macOS 10.13 or later changes how FileVault encryption keys
are generated. In previous versions of macOS on CoreStorage volumes, the keys used
in the FileVault encryption process were created when a user or organization turned on
FileVault on a Mac. In macOS on APFS volumes, the keys are generated either during user
creation, setting the first user’s password, or during the first login by a user of the Mac.
This implementation of the encryption keys, when they’re generated, and how they’re
stored are all part of a feature known as Secure Token. Specifically, a secure token is a
wrapped version of a key encryption key (KEK) protected by a userʼs password.
When deploying FileVault on APFS, the user can continue to:
• Use existing tools and processes, such as a personal recovery key (PRK) that can be
stored with a mobile device management (MDM) solution for escrow
• Create and use an institutional recovery key (IRK)
• Defer enablement of FileVault until a user logs in to or out of the Mac
In macOS 11, setting the initial password for the very first user on the Mac results in
that user being granted a secure token. In some workflows, that may not be the desired
behavior, as previously, granting the first secure token would have required the user
account to log in. To prevent this from happening, add ;DisabledTags;SecureToken to
the programmatically created user’s AuthenticationAuthority attribute prior to setting
the user’s password, as shown below:
sudo dscl . append /Users/<user name> AuthenticationAuthority

";DisabledTags;SecureToken"

Using Bootstrap Token
macOS 10.15 introduced a new feature—Bootstrap Token—to help with granting a secure
token to both mobile accounts and the optional device enrollment-created administrator
account (“managed administrator”). In macOS 11, the Bootstrap Token can grant a secure
token to any user logging in to a Mac computer, including local user accounts. Using the
Bootstrap Token feature of macOS 10.15 or later requires:
• Mac enrollment in MDM using Apple School Manager or Apple Business Manager, which
makes the Mac supervised
• MDM vendor support
In macOS 10.15.4 or later, a Bootstrap Token is generated and escrowed to MDM on the
first login by any user who is Secure Token–enabled if the MDM solution supports the
feature. A bootstrap token can also be generated and escrowed to MDM using the profiles
command-line tool, if needed.
In macOS 11, the bootstrap token may also be used for more than just granting secure
token to user accounts. On a Mac with Apple silicon, the bootstrap token, if available, can
be used to authorize the installation of both kernel extensions and software updates when
managed using MDM.
How Apple protects users’ personal data

Protecting app access to user data
In addition to encrypting data at rest, Apple devices help prevent apps from accessing a
user’s personal information without permission using various technologies including Data
Vault. In Settings in iOS and iPadOS, or System Preferences in macOS, users can see which
apps they have permitted to access certain information as well as grant or revoke any
future access. Access is enforced in the following:
• iOS, iPadOS, and macOS: Calendars, Camera, Contacts, Microphone, Photos,

Reminders, Speech recognition
• iOS and iPadOS: Bluetooth, Home, Media, Media apps and Apple Music, Motion and
fitness
• iOS and watchOS: Health
• macOS: Input monitoring (for example, keyboard strokes), Prompt, Screen recording
(for example, static screen shots and video), System Preferences
In iOS 13.4 or later and (iPadOS 13.4) or later, all third-party apps automatically have their
data protected in a Data Vault. Data Vault helps protect against unauthorized access to the
data even from processes that aren’t themselves sandboxed.
If the user signs in to iCloud, apps in iOS and iPadOS are granted access by default to
iCloud Drive. Users may control each app’s access under iCloud in Settings. iOS and
iPadOS also provide restrictions that prevent data movement between apps and accounts
installed by a mobile device management (MDM) solution and those installed by the user.

Protecting access to user’s health data
HealthKit provides a central repository for health and fitness data on iPhone and
Apple Watch. HealthKit also works directly with health and fitness devices, such as
compatible Bluetooth Low Energy (BLE) heart rate monitors and the motion coprocessor
built into many iOS devices. All HealthKit interaction with health and fitness apps,
healthcare institutions, and health and fitness devices require permission of the user. This
data is stored in the Data Protection class Protected Unless Open. Access to the data is
relinquished 10 minutes after the device locks, and data becomes accessible the next time
user enters their passcode or uses Touch ID or Face ID to unlock the device.
Collecting and storing health and fitness data

HealthKit also collects and stores management data, such as access permissions for apps,
names of devices connected to HealthKit, and scheduling information used to launch apps
when new data is available. This data is stored in the Data Protection class Protected Until
First User Authentication. Temporary journal files store health records that are generated
when the device is locked, such as when the user is exercising. These are stored in the
Data Protection class Protected Unless Open. When the device is unlocked, the temporary
journal files are imported into the primary health databases, then deleted when the merge
is completed.
Health data can be stored in iCloud. End-to-end encryption for Health data requires iOS 12
or later and two-factor authentication. Otherwise, the user’s data is still encrypted in
storage and transmission but isn’t encrypted end-to-end. After the user turns on two-
factor authentication and updates to iOS 12 or later, the user’s health data is migrated to
end-to-end encryption.
If the user backs up their device using iTunes (in macOS 10.14 or earlier) or the Finder
(macOS 10.15 or later), health data is stored only if the backup is encrypted.
Clinical health records

Users can sign in to supported health systems within the Health app to obtain a copy
of their clinical health records. When connecting a user to a health system, the user
authenticates using OAuth 2 client credentials. After connecting, clinical health record data
is downloaded directly from the health institution using a TLS 1.3 protected connection.
Once downloaded, clinical health records are securely stored alongside other health data.
Health data integrity

Data stored in the database includes metadata to track the provenance of each data
record. This metadata includes an app identifier that identifies which app stored the record.
Additionally, an optional metadata item can contain a digitally signed copy of the record.
This is intended to provide data integrity for records generated by a trusted device. The
format used for the digital signature is the Cryptographic Message Syntax (CMS) specified
in RFC 5652.

Health data access by third-party apps
Access to the HealthKit API is controlled with entitlements, and apps must conform to
restrictions about how the data is used. For example, apps aren’t allowed to use health
data for advertising. Apps are also required to provide users with a privacy policy that
details its use of health data.
Access to health data by apps is controlled by the user’s Privacy settings. Users are asked
to grant access when apps request access to health data, similar to Contacts, Photos,
and other iOS data sources. However, with health data, apps are granted separate access
for reading and writing data, as well as separate access for each type of health data.
Users can view, and revoke, permissions they’ve granted for accessing health data under
Settings > Health > Data Access & Devices.
If granted permission to write data, apps can also read the data they write. If granted the
permission to read data, they can read data written by all sources. However, apps can’t
determine access granted to other apps. In addition, apps can’t conclusively tell if they
have been granted read access to health data. When an app doesn’t have read access, all
queries return no data—the same response that an empty database would return. This
prevents apps from inferring the user’s health status by learning which types of data the
user is tracking.
Medical ID for users

The Health app gives users the option of filling out a Medical ID form with information that
could be important during a medical emergency. The information is entered or updated
manually and isn’t synced with the information in the health databases.
The Medical ID information is viewed by tapping the Emergency button on the Lock Screen.
The information is stored on the device using the Data Protection class No Protection so
that it’s accessible without having to enter the device passcode. Medical ID is an optional
feature that enables users to decide how to balance both safety and privacy concerns.
This data is backed up in iCloud Backup in iOS 13 or earlier. In iOS 14, Medical ID is synced
between devices using CloudKit and has the same encryption characteristics as the rest of
health data.
Digital signing and encryption

Access control lists
Keychain data is partitioned and protected with access control lists (ACLs). As a result,
credentials stored by third-party apps can’t be accessed by apps with different identities
unless the user explicitly approves them. This protection provides a mechanism for
securing authentication credentials in Apple devices across a range of apps and services
within the organization.

Mail
In the Mail app, users can send messages that are digitally signed and encrypted. Mail
automatically discovers appropriate RFC 5322 case-sensitive email address subject
or subject alternative names on digital signing and encryption certificates on attached
Personal Identification Verification (PIV) tokens in compatible smart cards. If a configured
email account matches an email address on a digital signing or encryption certificate on an
attached PIV token, Mail automatically displays the signing button in the toolbar of a new
message window. If Mail has the recipient’s email encryption certificate or can discover it
in the Microsoft Exchange global address list (GAL), an unlocked icon appears in the new
message toolbar. A locked lock icon indicates the message will be sent encrypted with the
recipient’s public key.
Per-message S/MIME
iOS, iPadOS, and macOS support per-message S/MIME. This means that S/MIME users can
choose to always sign and encrypt messages by default or to selectively sign and encrypt
individual messages.
Identities used with S/MIME can be delivered to Apple devices using a configuration profile,
a mobile device management (MDM) solution, the Simple Certificate Enrollment Protocol
(SCEP), or Microsoft Active Directory Certificate Authority.
Smart cards
macOS 10.12 or later includes native support for PIV cards. These cards are widely used
in commercial and government organizations for two-factor authentication, digital signing,
and encryption.
Smart cards include one or more digital identities that have a pair of public and private
keys and an associated certificate. Unlocking a smart card with the personal identification
number (PIN) provides access to the private keys used for authentication, encryption, and
signing operations. The certificate determines what a key can be used for, what attributes
are associated with it, and whether it’s validated (signed) by a certificate authority (CA)
certificate.
Smart cards can be used for two-factor authentication. The two factors needed to unlock
a card are “something the user has” (the card) and “something the user knows” (the PIN).
macOS 10.12 or later also has native support for smart card Login Window authentication
and client certificate authentication to websites on Safari. It also supports Kerberos
authentication using key pairs (PKINIT) for single sign-on to Kerberos-supported services.
To learn more about smart cards and macOS, see Intro to smart card integration in the
Deployment Reference for Mac.

Encrypted disk images
In macOS, encrypted disk images serve as secure containers in which users can store or
transfer sensitive documents and other files. Encrypted disk images are created using Disk
Utility, located in /Applications/Utilities/. Disk images can be encrypted using either 128-
bit or 256-bit AES encryption. Because a mounted disk image is treated as a local volume
connected to a Mac, users can copy, move, and open files and folders stored in it. As with
FileVault, the contents of a disk image are encrypted and decrypted in real time. With
encrypted disk images, users can safely exchange documents, files, and folders by saving
an encrypted disk image to removable media, sending it as a mail message attachment, or
storing it on a remote server. For more information on encrypted disk images, see the
Disk Utility User Guide.

App security
App security overview

Today, apps are among the most critical elements of a security architecture. Even as apps
provide amazing productivity benefits for users, they also have the potential to negatively
impact system security, stability, and user data if they’re not handled properly.
Because of this, Apple provides layers of protection to help ensure that apps are free
of known malware and haven’t been tampered with. Additional protections enforce that
access from apps to user data is carefully mediated. These security controls provide a
stable, secure platform for apps, enabling thousands of developers to deliver hundreds
of thousands of apps for iOS, iPadOS, and macOS—all without impacting system integrity.
And users can access these apps on their Apple devices without undue fear of viruses,
malware, or unauthorized attacks.
On iPhone, iPad, and iPod touch, all apps are obtained from the App Store—and all apps
are sandboxed—to provide the tightest controls.
On Mac, many apps are obtained from the App Store, but Mac users also download and use
apps from the internet. To safely support internet downloading, macOS layers additional
controls. First, by default in macOS 10.15 or later, all Mac apps need to be notarized
by Apple to launch. This requirement helps to ensure that these apps are free of known
malware without requiring that the apps be provided through the App Store. In addition,
macOS includes state-of-the-art antivirus protection to block—and if necessary remove—
malware.
As an additional control across platforms, sandboxing helps protect user data from
unauthorized access by apps. And in macOS, data in critical areas is itself protected—
which helps ensure that users remain in control of access to files in Desktop, Documents,
Downloads, and other areas from all apps, whether the apps attempting access are
themselves sandboxed or not.

Native capability Third-party equivalent
Plug-in unapproved list, Safari extension unapproved Virus/Malware definitions

list
File quarantine Virus/Malware definitions
XProtect/YARA signatures Virus/Malware definitions
MRT (Malware Removal Tool) Endpoint protection
Gatekeeper Endpoint protection; enforces code signing on apps to

ensure only trusted software runs.
eficheck Endpoint protection; rootkit detection
(Necessary for a Mac without an Apple T2 Security

Chip)
Application firewall Endpoint protection; firewalling
Packet Filter (pf) Firewall solutions
System Integrity Protection Built into macO S
Mandatory Access Controls Built into macO S
Kext exclude list Built into macO S
Mandatory app code signing Built into macO S
App notarization Built into macO S
App security in iOS and iPadOS

App security overview for iOS and iPadOS
Unlike other mobile platforms, iOS and iPadOS don’t allow users to install potentially
malicious unsigned apps from websites or run untrusted apps. At runtime, code signature
checks that all executable memory pages are made as they are loaded to ensure that an
app hasn’t been modified since it was installed or last updated.
After an app is verified to be from an approved source, iOS and iPadOS enforce security
measures designed to prevent it from compromising other apps or the rest of the system.
App code signing process in iOS and iPadOS
Mandatory code signing

After the iOS or iPadOS kernel has started, it controls which user processes and apps can
be run. To ensure that all apps come from a known and approved source and haven’t been
tampered with, iOS and iPadOS require that all executable code be signed using an Apple-
issued certificate. Apps provided with the device, like Mail and Safari, are signed by Apple.
Third-party apps must also be validated and signed using an Apple-issued certificate.
Mandatory code signing extends the concept of chain of trust from the operating system
to apps and prevents third-party apps from loading unsigned code resources or using self-
modifying code.

How developers sign their apps
Developers can sign their apps through certificate validation (through the Apple Developer
Program). They can also embed frameworks inside their apps and have that code validated
with an Apple-issued certificate (through a team identifier string).
• Certificate validation: To develop and install apps in iOS or iPadOS devices, developers
must register with Apple and join the Apple Developer Program. The real-world
identity of each developer, whether an individual or a business, is verified by Apple
before their certificate is issued. This certificate enables developers to sign apps and
submit them to the App Store for distribution. As a result, all apps in the App Store
have been submitted by an identifiable person or organization, serving as a deterrent
to the creation of malicious apps. They have also been reviewed by Apple to ensure
they generally operate as described and don’t contain obvious bugs or other notable
problems. In addition to the technology already discussed, this curation process gives
users confidence in the quality of the apps they buy.
• Code signature validation: iOS and iPadOS allow developers to embed frameworks
inside of their apps, which can be used by the app itself or by extensions embedded
within the app. To protect the system and other apps from loading third-party code
inside of their address space, the system performs a code signature validation of all
the dynamic libraries that a process links against at launch time. This verification
is accomplished through the team identifier (Team ID), which is extracted from an
Apple-issued certificate. A team identifier is a 10-character alphanumeric string—for
example, 1A2B3C4D5F. A program may link against any platform library that ships with
the system or any library with the same team identifier in its code signature as the main
executable. Since the executables shipping as part of the system don’t have a team
identifier, they can only link against libraries that ship with the system itself.
Verifying enterprise apps

Businesses also have the ability to write in-house apps for use within their organization
and distribute them to their employees. Businesses and organizations can apply to the
Apple Developer Enterprise Program (ADEP) with a D-U-N-S number. Apple approves
applicants after verifying their identity and eligibility. After an organization becomes a
member of ADEP, it can register to obtain a provisioning profile that permits in-house apps
to run on devices it authorizes.
Users must have the provisioning profile installed to run the in-house apps. This ensures
that only the organization’s intended users are able to load the apps onto their iOS and
iPadOS devices. Apps installed through mobile device management (MDM) are implicitly
trusted because the relationship between the organization and the device is already
established. Otherwise, users have to approve the app’s provisioning profile in Settings.
Organizations can restrict users from approving apps from unknown developers. On first
launch of any enterprise app, the device must receive positive confirmation from Apple
that the app is allowed to run.

Security of runtime process in iOS and iPadOS
Sandboxing
All third-party apps are “sandboxed,” so they are restricted from accessing files stored
by other apps or from making changes to the device. Sandboxing prevents apps from
gathering or modifying information stored by other apps. Each app has a unique home
directory for its files, which is randomly assigned when the app is installed. If a third-party
app needs to access information other than its own, it does so only by using services
explicitly provided by iOS and iPadOS.
System files and resources are also shielded from the users’ apps. Most iOS and iPadOS
system files and resources run as the nonprivileged user “mobile,” as do all third-party
apps. The entire operating system partition is mounted as read-only. Unnecessary tools,
such as remote login services, aren’t included in the system software, and APIs don’t allow
apps to escalate their own privileges to modify other apps or iOS and iPadOS.
Use of entitlements
Access by third-party apps to user information, and to features such as iCloud and
extensibility, is controlled using declared entitlements. Entitlements are key-value pairs
that are signed in to an app and allow authentication beyond runtime factors, like UNIX
user ID. Since entitlements are digitally signed, they can’t be changed. Entitlements are
used extensively by system apps and daemons to perform specific privileged operations
that would otherwise require the process to run as root. This greatly reduces the potential
for privilege escalation by a compromised system app or daemon.
In addition, apps can only perform background processing through system-provided APIs.
This enables apps to continue to function without degrading performance or dramatically
impacting battery life.
Address Space Layout Randomization

Address Space Layout Randomization (ASLR) protects against the exploitation of
memory corruption bugs. Built-in apps use ASLR to ensure that all memory regions are
randomized upon launch. Randomly arranging the memory addresses of executable code,
system libraries, and related programming constructs reduces the likelihood of many
sophisticated exploits. For example, a return-to-libc attack attempts to trick a device into
executing malicious code by manipulating memory addresses of the stack and system
libraries. Randomizing the placement of these makes the attack far more difficult to
execute, especially across multiple devices. Xcode, and the iOS or iPadOS development
environments, automatically compile third-party programs with ASLR support turned on.
Execute Never feature

Further protection is provided by iOS and iPadOS using ARM’s Execute Never (XN) feature,
which marks memory pages as nonexecutable. Memory pages marked as both writable
and executable can be used only by apps under tightly controlled conditions: The kernel
checks for the presence of the Apple-only dynamic code-signing entitlement. Even then,
only a single mmap call can be made to request an executable and writable page, which is
given a randomized address. Safari uses this functionality for its JavaScript JIT compiler.

Supporting extensions in iOS, iPadOS, and macOS
iOS, iPadOS, and macOS allow apps to provide functionality to other apps by providing
extensions. Extensions are special-purpose signed executable binaries packaged within
an app. During installation, the system automatically detects extensions and makes them
available to other apps using a matching system.
Extension points
A system area that supports extensions is called an extension point. Each extension point
provides APIs and enforces policies for that area. The system determines which extensions
are available based on extension point–specific matching rules. The system automatically
launches extension processes as needed and manages their lifetime. Entitlements can be
used to restrict extension availability to particular system apps. For example, a Today view
widget appears only in Notification Center, and a sharing extension is available only from
the Sharing pane. Examples of extension points are Today widgets, Share, Actions, Photo
Editing, File Provider, and Custom Keyboard.
How extensions communicate

Extensions run in their own address space. Communication between the extension and
the app from which it was activated uses interprocess communications mediated by
the system framework. They don’t have access to each other’s files or memory spaces.
Extensions are designed to be isolated from each other, from their containing apps, and
from the apps that use them. They are sandboxed like any other third-party app and have
a container separate from the containing app’s container. However, they share the same
access to privacy controls as the container app. So if a user grants Contacts access to an
app, this grant is extended to the extensions that are embedded within the app but not to
the extensions activated by the app.
How custom keyboards are used

Custom keyboards are a special type of extension because it’s enabled by the user for the
entire system. After it’s enabled, a keyboard extension is used for any text field except
the passcode input and any secure text view. To restrict the transfer of user data, custom
keyboards run by default in a very restrictive sandbox that blocks access to the network,
to services that perform network operations on behalf of a process, and to APIs that would
allow the extension to exfiltrate typing data. Developers of custom keyboards can request
that their extension have Open Access, which lets the system run the extension in the
default sandbox after getting consent from the user.
MDM and extensions

For devices enrolled in a mobile device management (MDM) solution, document and
keyboard extensions obey Managed Open In rules. For example, the MDM solution
can prevent a user from exporting a document from a managed app to an unmanaged
Document Provider, or using an unmanaged keyboard with a managed app. Additionally,
app developers can prevent the use of third-party keyboard extensions within their app.

App protection and app groups in iOS and iPadOS
Adopting Data Protection in apps

The iOS Software Development Kit (SDK) for iOS and iPadOS offers a full suite of APIs that
make it easy for third-party and in-house developers to adopt Data Protection and help
ensure the highest level of protection in their apps. Data Protection is available for file and
database APIs, including NSFileManager, CoreData, NSData, and SQLite.
The Mail app database (including attachments), managed books, Safari bookmarks, app
launch images, and location data are also stored through encryption, with keys protected
by the user’s passcode on their device. Calendar (excluding attachments), Contacts,
Reminders, Notes, Messages, and Photos implement the Data Protection entitlement
Protected Until First User Authentication.
User-installed apps that don’t opt in to a specific Data Protection class receive Protected
Until First User Authentication by default.
Joining an App Group

Apps and extensions owned by a given developer account can share content when
configured to be part of an App Group. It’s up to the developer to create the appropriate
groups on the Apple Developer Portal and include the desired set of apps and extensions.
Once configured to be part of an App Group, apps have access to the following:
• A shared on-volume container for storage, which stays on the device as long as at least
one app from the group is installed
• Shared preferences
• Shared keychain items
The Apple Developer Portal ensures that App group IDs (GID) are unique across the app
ecosystem.
Verifying accessories in iOS and iPadOS

The Made for iPhone, iPad, and iPod touch (MFi) licensing program provides vetted
accessory manufacturers access to the iPod Accessories Protocol (iAP) and the necessary
supporting hardware components.
When an MFi accessory communicates with an iOS or iPadOS device using a Lightning
connector or through Bluetooth, the device asks the accessory to prove it’s been
authorized by Apple by responding with an Apple-provided certificate, which is verified by
the device. The device then sends a challenge, which the accessory must answer with a
signed response. This process is entirely handled by a custom integrated circuit (IC) that
Apple provides to approved accessory manufacturers and is transparent to the accessory
itself.
Accessories can request access to different transport methods and functionality—for

example, access to digital audio streams over the Lightning cable, or location information
provided over Bluetooth. An authentication IC ensures that only approved accessories
are granted full access to the device. If an accessory doesn’t support authentication,
its access is limited to analog audio and a small subset of serial (UART) audio playback
controls.

AirPlay also uses the authentication IC to verify that receivers have been approved by
Apple. AirPlay audio and CarPlay video streams use the MFi-SAP (Secure Association
Protocol), which encrypts communication between the accessory and device using
AES128 in counter (CTR) mode. Ephemeral keys are exchanged using ECDH key exchange
(Curve25519) and signed using the authentication IC’s 1024-bit RSA key as part of the
Station-to-Station (STS) protocol.
App security in macOS

App security overview in macOS
App security in macOS consists of a number of overlapping layers—the first of which is
the option to run only signed and trusted apps from the App Store. In addition, macOS
layers protections to ensure that apps downloaded from the internet are free of known
malware. macOS offers technologies to detect and remove malware, and offers additional
protections designed to prevent untrusted apps from accessing user data. Services
from Apple such as Notarization and XProtect and MRT updates are designed to prevent
malware installation and, when necessary, to provide for a quick and efficient detect-and-
respond process to block and remove any malware that may have at first avoided detection.
Ultimately, macOS users are free to operate within the security model that makes sense for
them—including running completely unsigned and untrusted code.
App code signing process in macOS

All apps from the App Store are signed by Apple to ensure that they haven’t been tampered
with or altered. Apple signs any apps provided with Apple devices.
In macOS 10.15, all apps distributed outside the App Store must be signed by the developer
using an Apple-issued Developer ID certificate (combined with a private key) and notarized
by Apple to run under the default Gatekeeper settings. Apps developed in-house should
also be signed with an Apple-issued Developer ID so that users can validate their integrity.
In macOS, code signing and notarization work independently—and can be performed by

different actors—for different goals. Code signing is performed by the developer using
their Developer ID certificate (issued by Apple), and verification of this signature proves
to the user that a developer’s software hasn’t been tampered with since the developer
built and signed it. Notarization can be performed by anyone in the software distribution
chain and proves that Apple has been provided a copy of the code to check for malware
and no known malware was found. The output of Notarization is a ticket, which is stored on
Apple servers and can be optionally stapled to the app (by anyone) without invalidating the
signature of the developer.
Mandatory Access Controls (MACs) require code signing to enable entitlements protected
by the system. For example, apps requiring access through the firewall must be code
signed with the appropriate MAC entitlement.

Gatekeeper and runtime protection in macOS
Gatekeeper
macOS includes a technology called Gatekeeper, which ensures that, by default, only
trusted software runs on a user’s Mac. When a user downloads and opens an app, a
plug-in, or an installer package from outside the App Store, Gatekeeper verifies that
the software is from an identified developer, is notarized by Apple to be free of known
malicious content, and hasn’t been altered. Gatekeeper also requests user approval before
opening downloaded software for the first time to make sure the user hasn’t been tricked
into running executable code they believed to simply be a data file.
By default, Gatekeeper ensures that all downloaded software has been signed by the App
Store or signed by a registered developer and notarized by Apple. Both the App Store
review process and the notarization pipeline ensure that apps contain no known malware.
Therefore, by default all software in macOS is checked for known malicious content the
first time it’s opened, regardless of how it arrived on the Mac.
Users and organizations have the option to allow only software installed from the App
Store. Alternatively, users can override Gatekeeper policies to open any software unless
restricted by a mobile device management (MDM) solution. Organizations can use MDM to
configure Gatekeeper settings, including allowing software signed with alternate identities.
Gatekeeper can also be completely disabled, if necessary.
Gatekeeper also protects against the distribution of malicious plug-ins with benign apps.
Here, using the app triggers loading a malicious plug-in without the user’s knowledge.
When necessary, Gatekeeper opens apps from randomized, read-only locations, preventing
the automatic loading of plug-ins distributed alongside the app.
Runtime protection
System files, resources, and the kernel are shielded from a user’s app space. All apps from
the App Store are sandboxed to restrict access to data stored by other apps. If an app
from the App Store needs to access data from another app, it can do so only by using the
APIs and services provided by macOS.
Protecting against malware in macOS

Apple operates a threat intelligence process to quickly identify and block malware. Malware
defenses are structured in three layers:
1. Prevent launch or execution of malware: App Store or Gatekeeper and Notarization
2. Block malware from running on customer systems: Gatekeeper, Notarization, and

XProtect
3. Remediate malware that has executed: MRT
The first layer of defense is designed to inhibit the distribution of malware, and prevent it
from launching even once—this is the goal of the App Store, and Gatekeeper combined
with Notarization.
The next layer of defense is to ensure that if malware appears on any Mac, it’s quickly
identified and blocked, both to halt spread and to remediate the Mac systems it’s
already gained a foothold on. XProtect adds to this defense, along with Gatekeeper and
Notarization.

Finally, MRT acts to remediate malware that has managed to successfully execute.
These protections combine to support best-practice protection from viruses and malware.
There are additional protections, particularly on a Mac with Apple silicon, to limit the
potential damage of malware that does manage to execute. See Protecting app access to
user data for ways that macOS can help protect user data from malware, and Operating
system integrity for ways macOS can limit the actions malware can take on the system.
Notarization
Notarization is a malware scanning service provided by Apple. Developers who want to
distribute apps for macOS outside the App Store submit their apps for scanning as part
of the distribution process. Apple scans this software for known malware and, if none is
found, issues a Notarization ticket. Typically, developers staple this ticket to their app so
Gatekeeper can verify and launch the app, even offline.
Apple can also issue a revocation ticket for apps known to be malicious—even if they’ve
been previously notarized. macOS regularly checks for new revocation tickets so that
Gatekeeper has the latest information and can block launch of such files. This process
can very quickly block malicious apps because updates happen in the background much
more frequently than even the background updates that push new XProtect signatures. In
addition, this protection can be applied to both apps that have been previously and those
that haven’t.
XProtect
macOS includes built-in antivirus technology called XProtect for the signature-based
detection of malware. The system uses YARA signatures, a tool used to conduct signature-
based detection of malware, which Apple updates regularly. Apple monitors for new
malware infections and strains, and updates signatures automatically—independent from
system updates—to help defend a Mac from malware infections. XProtect automatically
detects and blocks the execution of known malware. In macOS 10.15 or later, XProtect
checks for known malicious content whenever:
• An app is first launched
• An app has been changed (in the file system)
• XProtect signatures are updated
When XProtect detects known malware, the software is blocked and the user is notified
and given the option to move the software to the Trash.
Note: Notarization is effective against known files (or file hashes) and can be used on
apps that have been previously launched. The signature-based rules of XProtect are more
generic than a specific file hash, so it can find variants that Apple has not seen. XProtect
operates with a slower update cycle than notarization and scans only apps that have been
changed or at first launch.

Malware Removal Tool
Should malware make its way onto a Mac, macOS also includes technology to remediate
infections. The Malware Removal Tool (MRT) is an engine in macOS that remediates
infections based on updates automatically delivered from Apple (as part of automatic
updates of system data files and security updates). MRT removes malware upon receiving
updated information, and it continues to check for infections on restart and login. MRT
doesn’t automatically reboot the Mac.
Automatic security updates

Apple issues the updates for XProtect and MRT automatically based on the latest threat
intelligence available. By default, macOS checks for these updates daily. Notarization
updates are distributed using CloudKit sync and are much more frequent.
Response process
When new malware is discovered, a number of steps may be performed:
• Any associated Developer ID certificates are revoked.
• Notarization revocation tickets are issued for all files (apps and associated files).
• XProtect signatures are developed and released.
• MRT signatures are developed and released.
• These signatures are also applied retroactively to previously notarized software, and
any new detections can result in one or more of the previous actions occurring.
Ultimately, a malware detection launches a series of steps over the next seconds, hours,
and days that follow to propagate the best protections possible to Mac users.
Controlling app access to files in macOS

Apple believes that users should have full transparency, consent, and control over what
apps are doing with their data. In macOS 10.15, this model is enforced by the system
to ensure that all apps must obtain user consent before accessing files in Documents,
Downloads, Desktop, iCloud Drive, or network volumes. In macOS 10.13 or later, apps that
require access to the full storage device must be explicitly added in System Preferences.
In addition, accessibility and automation capabilities require user permission to ensure
they don’t circumvent other protections. Depending on the access policy, users may
be prompted or must change the setting in System Preferences > Security & Privacy >
Privacy:
Item User prompted by app User must edit system privacy

settings
Accessibility
Full internal storage access
Files and folders
Note: Includes: Desktop,

Documents, Downloads, network
volumes, and removable volumes
Automation (Apple events)

Items in the user’s Trash are protected from any apps that are using Full Disk Access; the
user won’t get prompted for app access. If the user wants apps to access the files, they
must be moved from the Trash to another location.
A user who enables FileVault on a Mac is asked to provide valid credentials before
continuing the boot process and gain access to specialized startup modes. Without valid
login credentials or a recovery key, the entire volume remains encrypted and is protected
from unauthorized access, even if the physical storage device is removed and connected
to another computer.
To protect data in an enterprise setting, IT should define and enforce FileVault

configuration policies using mobile device management (MDM). Organizations have several
options for managing encrypted volumes, including institutional recovery keys, personal
recovery keys (that can optionally be stored with MDM for escrow), or a combination of
both. Key rotation can also be set as a policy in MDM.
Secure features in the Notes app

Secure notes
The Notes app includes a secure notes feature that allows users to protect the contents of
specific notes. Secure notes are end-to-end encrypted using a user-provided passphrase
that is required to view the notes on iOS, iPadOS, macOS devices, and the iCloud website.
Each iCloud account (including “On my” device accounts) can have a separate passphrase.
When a user secures a note, a 16-byte key is derived from the user’s passphrase using
PBKDF2 and SHA256. The note and all of its attachments are encrypted using AES with
Galois/Counter Mode (AES-GCM). New records are created in Core Data and CloudKit to
store the encrypted note, attachments, tag, and initialization vector. After the new records
are created, the original unencrypted data is deleted. Attachments that support encryption
include images, sketches, tables, maps, and websites. Notes containing other types of
attachments can’t be encrypted, and unsupported attachments can’t be added to secure
notes.
To view a secure note, the user must enter their passphrase or authenticate using Touch ID
or Face ID. After successfully authenticating the user, whether to view or create a secure
note, Notes opens a secure session. While the secure session is open, the user can view or
secure other notes without additional authentication. However, the secure session applies
only to notes protected with the provided passphrase. The user still needs to authenticate
for notes protected by a different passphrase. The secure session is closed when:
• The user taps the Lock Now button in Notes
• Notes is switched to the background for more than 3 minutes (8 minutes in macOS)
• The iOS or iPadOS device locks
To change the passphrase on a secure note, the user must enter the current passphrase,
because Touch ID and Face ID aren’t available when changing the passphrase. After
choosing a new passphrase, the Notes app rewraps, in the same account, the keys of all
existing notes that are encrypted by the previous passphrase.

If a user mistypes the passphrase three times in a row, Notes shows a user-supplied hint if
one was provided by the user at setup. If the user still doesn’t remember their passphrase,
they can reset it in Notes settings. This feature allows users to create new secure notes
with a new passphrase, but it won’t allow them to see previously secured notes. The
previously secured notes can still be viewed if the old passphrase is remembered.
Resetting the passphrase requires the user’s iCloud account passphrase.
Shared notes
Notes that aren’t end-to-end encrypted with a passphrase can be shared with others.
Shared notes still use the CloudKit encrypted data type for any text or attachments that
the user puts in a note. Assets are always encrypted with a key that’s encrypted in the
CKRecord. Metadata, such as the creation and modification dates, aren’t encrypted.
CloudKit manages the process by which participants can encrypt and decrypt each other’s
data.
Secure features in the Shortcuts app

In the Shortcuts app, shortcuts are optionally synced across Apple devices using iCloud.
Shortcuts can also be shared with other users through iCloud. Shortcuts are stored locally
in an encrypted format.
Custom shortcuts are versatile—they’re similar to scripts or programs. When downloading

shortcuts from the internet, the user is warned that the shortcut hasn’t been reviewed by
Apple and is given the opportunity to inspect the shortcut. To protect against malicious
shortcuts, updated malware definitions are downloaded to identify malicious shortcuts at
runtime.
Custom shortcuts can also run user-specified JavaScript on websites in Safari when
invoked from the share sheet. To protect against malicious JavaScript that, for example,
tricks the user into running a script on a social media website that harvests their data,
the JavaScript is validated against the aforementioned malware definitions. The first time
a user runs JavaScript on a domain, the user is prompted to allow shortcuts containing
JavaScript to run on the current webpage for that domain.

Services security
Services security overview

Apple has built a robust set of services to help users get even more utility and productivity
out of their devices. These services provide powerful capabilities for cloud storage, sync,
password storage, authentication, payment, messaging, communications, and more, all
while protecting users’ privacy and the security of their data.
These services include iCloud, Sign in with Apple, Apple Pay, iMessage, Business Chat,
FaceTime, Find My, and Continuity and may require an Apple ID or Managed Apple ID. In
some cases, a Managed Apple ID can’t be used with a specific service, like Apple Pay.
Note: Not all Apple services and content are available in all countries or regions.
Apple ID and Managed Apple ID

Apple ID security overview
Overview
An Apple ID is the account used to sign in to Apple services such as iCloud, iMessage,
FaceTime, the iTunes Store, the App Store, the Apple TV app, Apple Books, and more. It’s
important for users to keep their Apple IDs secure to prevent unauthorized access to their
accounts. To help with this, Apple IDs require strong passwords that:
• Must be at least eight characters in length
• Must contain both letters and numbers
• Must not contain more than three consecutive identical characters
• Can’t be a commonly used password
Users are encouraged to exceed these guidelines by adding extra characters and
punctuation marks to make their passwords even stronger.
Apple also notifies users in email and/or push notifications when important changes are
made to their account—for example, if a password or billing information has been changed
or the Apple ID has been used to sign in on a new device. If anything looks unfamiliar, users
are instructed to change their Apple ID password immediately.

In addition, Apple employs a variety of policies and procedures designed to protect user
accounts. These include limiting the number of retries for sign-in and password reset
attempts, active fraud monitoring to help identify attacks as they occur, and regular policy
reviews that allow Apple to adapt to any new information that could affect user security.
Note: The Managed Apple ID password policy is set by an administrator in Apple School
Manager or Apple Business Manager.
Two-factor authentication
To help users further secure their accounts, Apple offers two-factor authentication—an
extra layer of security for Apple IDs. It’s designed to ensure that only the account’s owner
can access the account, even if someone else knows the password. With two-factor
authentication, a user’s account can be accessed on only trusted devices, such as the
user’s iPhone, iPad, iPod touch, or Mac, or on other devices after completing a verification
from one of these trusted devices or a trusted phone number. To sign in for the first time
on any new device, two pieces of information are required—the Apple ID password and a
six-digit verification code that’s displayed on the user’s trusted devices or sent to a trusted
phone number. By entering the code, the user confirms that they trust the new device and
that it’s safe to sign in. Because a password alone is no longer enough to access a user’s
account, two-factor authentication improves the security of the user’s Apple ID and all the
personal information they store with Apple. It’s integrated directly into iOS, iPadOS, macOS,
tvOS, watchOS, and the authentication systems used by Apple websites.
When a user signs in to an Apple website using a web browser, a second factor request is
sent to all trusted devices associated with the user’s iCloud account, requesting approval
of the web session. If the user is signing in to an Apple website from a browser on a trusted
device, they see the verification code displayed locally on the device they’re using. When
the user enters the code on that device, the web session is approved.
Account recovery
If an Apple ID account password is forgotten, a user can reset it on a trusted device. If a
trusted device isn’t available and the password is known, a user can use a trusted phone
number can be used to authenticate through SMS verification. In addition, to provide
immediate recovery for an Apple ID, a previously used passcode can be used to reset
in conjunction with SMS. If these options aren’t possible, the account recovery process
must be followed. For more information, see the Apple Support article Recover your
Apple ID when you can’t reset your password.
Managed Apple ID security

Managed Apple IDs function much like an Apple ID but are owned and controlled by
enterprise or educational organizations. These organizations can reset passwords, limit
purchasing and communications such as FaceTime and Messages, and set up role-based
permissions for employees, staff members, teachers, and students.
For Managed Apple IDs, some services are disabled (for example, Apple Pay, iCloud
Keychain , HomeKit, and Find My).

Inspecting Managed Apple IDs
Managed Apple IDs also support inspection, which allows organizations to comply with
legal and privacy regulations. An Apple School Manager administrator, manager, or teacher
can inspect specific Managed Apple ID accounts.
Inspectors can monitor only accounts that are below them in the organization’s hierarchy.
For example, teachers can monitor students, managers can inspect teachers and students,
and administrators can inspect managers, teachers, and students.
When inspecting credentials are requested using Apple School Manager, a special
account is issued that has access to only the Managed Apple ID for which inspecting was
requested. The inspector can then read and modify the user’s content stored in iCloud
or in CloudKit-enabled apps. Every request for auditing access is logged in Apple School
Manager. The logs show who the inspector was, the Managed Apple ID the inspector
requested access to, the time of the request, and whether the inspection was performed.
Managed Apple IDs and personal devices

Managed Apple IDs can also be used with personally owned iOS and iPadOS devices and
Mac computers. Students sign in to iCloud using the Managed Apple ID issued by the
institution and an additional home-use password, which serves as the second factor of the
Apple ID two-factor authentication process. While students are using a Managed Apple ID
on a personal device, iCloud Keychain isn’t available, and the institution might restrict
other features such as FaceTime or Messages. Any iCloud documents created by students
when they are signed in are subject to audit as described previously in this section.
iCloud
iCloud security overview
iCloud stores a user’s contacts, calendars, photos, documents, and more and keeps the
information up to date across all of their devices automatically. iCloud can also be used
by third-party apps to store and sync documents as well as key values for app data as
defined by the developer. Users set up iCloud by signing in with an Apple ID and choosing
which services they would like to use. Certain iCloud features, iCloud Drive, and iCloud
Backup can be disabled by IT administrators using mobile device management (MDM)
configuration profiles. The service is agnostic about what is being stored and handles all
file content the same way, as a collection of bytes.
Each file is broken into chunks and encrypted by iCloud using AES128 and a key derived
from each chunk’s contents, with the keys using SHA256. The keys and the file’s metadata
are stored by Apple in the user’s iCloud account. The encrypted chunks of the file are
stored, without any user-identifying information or the keys, using both Apple and third-
party storage services—such as Amazon Web Services or Google Cloud Platform—but
these partners don’t have the keys to decrypt the user’s data stored on their servers.

iCloud Drive security
iCloud Drive adds account-based keys to protect documents stored in iCloud. iCloud Drive
chunks and encrypts file contents and stores the encrypted chunks using third-party
services. However, the file content keys are wrapped by record keys stored with the iCloud
Drive metadata. These record keys are in turn protected by the user’s iCloud Drive service
key, which is then stored with the user’s iCloud account. Users get access to their iCloud
documents’ metadata by having authenticated with iCloud, but they must also possess the
iCloud Drive service key to expose protected parts of iCloud Drive storage.
iCloud Drive backup

iCloud also backs up information—including device settings, app data, photos, and videos
in the Camera Roll, and conversations in the Messages app—daily over Wi-Fi. iCloud
secures the content by encrypting it when it’s sent over the internet, storing it in an
encrypted format and using secure tokens for authentication. iCloud Backup occurs only
when the device is locked, connected to a power source, and has Wi-Fi access to the
internet. Because of the encryption used in iOS and iPadOS, iCloud Backup is designed to
keep data secure while allowing incremental, unattended backup and restoration to occur.
When files are created in Data Protection classes that aren’t accessible when the device
is locked, their per-file keys are encrypted, using the class keys from the iCloud Backup
keybag and backing the files up to iCloud in their original, encrypted state. All files are
encrypted during transport and, when stored, encrypted using account-based keys, as
described in CloudKit.
The iCloud Backup keybag contains asymmetric (Curve25519) keys for Data Protection
classes that aren’t accessible when the device is locked. The backup set is stored in the
user’s iCloud account and consists of a copy of the user’s files and the iCloud Backup
keybag. The iCloud Backup keybag is protected by a random key, which is also stored with
the backup set. (The user’s iCloud password isn’t used for encryption, so changing the
iCloud password won’t invalidate existing backups.)
While the user’s keychain database is backed up to iCloud, it remains protected by a UID-
tangled key. This allows the keychain to be restored only to the same device from which it
originated, and it means no one else, including Apple, can read the user’s keychain items.
On restore, the backed-up files, iCloud Backup keybag, and the key for the keybag are
retrieved from the user’s iCloud account. The iCloud Backup keybag is decrypted using
its key, then the per-file keys in the keybag are used to decrypt the files in the backup set,
which are written as new files to the file system, thus reencrypting them according to their
Data Protection class.

Security of iCloud Backup
The following content is backed up using iCloud Backup:
• Records for purchased music, movies, TV shows, apps, and books. A user’s iCloud
Backup includes information about purchased content present on the user’s device,
but not the purchased content itself. When the user restores from an iCloud Backup,
their purchased content is automatically downloaded from the iTunes Store, the App
Store, the Apple TV app, or Apple Books. Some types of content aren’t downloaded
automatically in all countries or regions, and previous purchases may be unavailable if
they have been refunded or are no longer available in the store. Full purchase history is
associated with a user’s Apple ID.
• Photos and videos on a user’s devices. Note that if a user turns on iCloud Photos in
iOS 8.1 or later, iPadOS 13.1 or later, or OS X 10.10.3 or later, their photos and videos are
already stored in iCloud, so they aren’t included in the user’s iCloud Backup.
• Contacts, calendar events, reminders, and notes
• Device settings
• App data
• Home screen and app organization
• HomeKit configuration
• Medical ID data
• Visual Voicemail password (requires the SIM card that was in use during backup)
• iMessage, Business Chat, text (SMS), and MMS messages (requires the SIM card that
was in use during backup)
When Messages in iCloud is enabled, iMessage, Business Chat, text (SMS), and MMS
messages are removed from the user’s existing iCloud Backup and are instead stored in an
end-to-end encrypted CloudKit container for Messages. The user’s iCloud Backup retains a
key to that container. If the user later disables iCloud Backup, that container’s key is rolled,
the new key is stored only in iCloud Keychain (inaccessible to Apple and any third parties),
and new data written to the container can’t be decrypted with the old container key.
The key used to restore the messages in iCloud Backup is placed in two locations, iCloud
Keychain and a backup in CloudKit. The backup in CloudKit is done if iCloud Backup is
enabled and unconditionally restored whether the user restores an iCloud backup or not.
CloudKit end-to-end encryption

Many Apple services, listed in the Apple Support article iCloud security overview, use
end-to-end encryption with a CloudKit service key protected by iCloud Keychain syncing.
For these CloudKit containers, the key hierarchy is rooted in iCloud Keychain and therefore
shares the security characteristics of iCloud Keychain—namely, the keys are available
only on the user’s trusted devices, and not to Apple or any third party. If access to iCloud
Keychain data is lost, the data in CloudKit is reset; and if data is available from the trusted
local device, it’s uploaded again to CloudKit. For more information, see Escrow security for
iCloud Keychain.

Messages in iCloud also uses CloudKit end-to-end encryption with a CloudKit service key
protected by iCloud Keychain syncing. If the user has enabled iCloud Backup, the CloudKit
service key used for the Messages in iCloud container is backed up to iCloud to allow
the user to recover their messages even if they have lost access to iCloud Keychain and
their trusted devices. This iCloud service key is rolled whenever the user turns off iCloud
Backup.
Situation User recovery options for CloudKit end-to-end

encryption
Access to trusted device Data recovery possible using a trusted device or iC loud
Keychain recovery.
No trusted devices Data recovery only possible using iCloud Keychain

recovery.
iC loud Backup enabled and access to trusted device Data recovery possible using iCloud Backup, access to
a trusted device, or iCloud Keychain recovery.
iC loud Backup enabled and no access to trusted device Data recovery possible using iCloud Backup or iC loud
Keychain recovery.
iC loud Backup disabled and access to trusted device Data recovery possible using a trusted device or iC loud
Keychain recovery.
Backup disabled and no trusted devices Data recovery only possible using iCloud Keychain
recovery.
Passcode and password management

Password security overview
iOS, iPadOS, and macOS make it easy for users to authenticate to third-party apps and
websites that use passwords. The best way to manage passwords is not to have to use
one. Sign in with Apple lets users sign in to third-party apps and websites without having
to create and manage an additional account or password while protecting the sign-in with
their two-factor authentication for Apple ID. For sites that don’t support Sign in with Apple,
the Automatic Strong Password feature enable a user’s devices to automatically create,
sync, and enter unique strong passwords for sites and apps. In iOS and iPadOS, passwords
are saved to a special Password AutoFill keychain that’s user controlled and manageable by
going to Settings > Passwords.
In macOS, saved passwords can be managed in Safari Passwords preferences. This sync
system can also be used to sync passwords that are manually created by the user.

Sign in with Apple security
Sign in with Apple is a privacy-friendly alternative to other single sign-on systems. It
provides the convenience and efficiency of one-tap sign-in while giving the user more
transparency and control over their personal information.
Sign in with Apple allows users to set up an account and sign in to apps and websites
using the Apple ID they already have, and it gives them more control over their personal
information. Apps can only ask for the user’s name and email address when setting up an
account, and the user always has a choice: They can share their personal email address
with an app or choose to keep their personal email private and use the new Apple private
email relay service instead. This email relay service shares a unique, anonymized email
address that forwards to the user’s personal address so they can still receive useful
communication from the developer while maintaining a degree of privacy and control over
their personal information.
Sign in with Apple is built for security. Every Sign in with Apple user is required to have
two-factor authentication enabled for their Apple ID. Two-factor authentication helps
secure not only the user’s Apple ID but also the accounts they establish with their apps.
Furthermore, Apple has developed and integrated a privacy-friendly antifraud signal into
Sign in with Apple. This signal gives developers confidence that the new users they acquire
are real people and not bots or scripted accounts.
Automatic strong passwords

When iCloud Keychain is enabled, iOS, iPadOS, and macOS create strong, random, unique
passwords when users sign up for or change their password on a website in Safari. In iOS
and iPadOS, automatic strong password generation is also available in apps. Users must
opt out of using strong passwords. Generated passwords are saved in the keychain and
kept up to date across devices with iCloud Keychain, when it’s enabled.
By default, passwords generated by iOS and iPadOS are 20 characters long. They contain
one digit, one uppercase character, two hyphens, and 16 lowercase characters. These
generated passwords are strong, containing 71 bits of entropy.
Passwords are generated based on heuristics that determine whether a password-field

experience is for password creation. If the heuristic fails to recognize a context-specific
password being used at password creation, app developers can set UITextContentType.
newPassword on their text field and web developers can set autocomplete= “new-
password” on their <input> elements.
To ensure that generated passwords are compatible with the relevant services,
apps and websites can provide rules. Developers provide these rules using
UITextInputPasswordRules or the passwordrules attribute on their <input> elements.
Devices then generate the strongest password they can that fulfills these rules.

Password AutoFill security
Password AutoFill automatically fills credentials stored in the keychain. The iCloud
Keychain password manager and Password AutoFill provide the following features:
• Filling in credentials in apps and websites
• Generating strong passwords
• Saving passwords in both apps and websites in Safari
• Sharing passwords securely to a users’ contacts
• Providing passwords to a nearby Apple TV that’s requesting credentials
Generating and saving passwords within apps, as well as providing passwords to Apple TV,
are available only in iOS and iPadOS.
Password AutoFill in apps

iOS and iPadOS allow users to input saved user names and passwords into credential-
related fields in apps, similar to the way Password AutoFill works in Safari. In iOS and
iPadOS, users tap a key affordance in the software keyboard’s QuickType bar. In macOS,
for apps built with Mac Catalyst, a Passwords drop-down menu appears below credential-
related fields.
When an app is strongly associated with a website that uses the same app-website
association mechanism and that’s powered by the same apple-app-site-association
file, the iOS and iPadOS QuickType bar and macOS drop-down menu directly suggest
credentials for the app, if any are saved to the Password AutoFill Keychain. This allows
users to choose to disclose Safari-saved credentials to apps with the same security
properties, without those apps having to adopt an API.
Password AutoFill exposes no credential information to an app until a user consents to

release a credential to the app. The credential lists are drawn from or presented out of the
app’s process.
When an app and website have a trusted relationship and a user submits credentials within
an app, iOS and iPadOS may prompt the user to save those credentials to the Password
AutoFill keychain for later use.
App access to saved passwords

iOS, iPadOS, and macOS apps can request the Password AutoFill keychain’s help with
signing a user in using ASAuthorizationPasswordProvider and SecAddSharedWebCredential.
The password provider and its request can be used in conjunction with Sign in with Apple,
so that the same API is called to help users sign into an app, regardless of whether the
user’s account is password based or was created using Sign in with Apple.
Apps can access saved passwords only if the app developer and website administrator
have given their approval and the user has given consent. App developers express their
intent to access Safari saved passwords by including an entitlement in their app. The
entitlement lists the fully qualified domain names of associated websites, and the websites
must place a file on their server listing the unique app identifiers of apps approved by
Apple.

When an app with the com.apple.developer.associated-domains entitlement is installed,
iOS and iPadOS make a TLS request to each listed website, requesting one of the following
files:
• apple-app-site-association
• .well-known/apple-app-site-association
If the file lists the app identifier of the app being installed, then iOS and iPadOS mark the
website and app as having a trusted relationship. Only with a trusted relationship will calls
to these two APIs result in a prompt to the user, who must agree before any passwords are
released to the app, updated, or deleted.
Password security recommendations
Overview
The Password AutoFill passwords list in iOS, iPadOS, and macOS indicates which of a
user’s saved passwords will be reused with other websites, passwords that are considered
weak, and passwords that have been compromised by a data leak.
Using the same password for more than one service may leave those accounts vulnerable
to a credential-stuffing attack. If a service is breached and passwords are leaked,
attackers may try the same credentials on other services to compromise additional
accounts. Passwords are marked as reused if the same password is seen used for more
than one saved password across different domains.
Passwords are marked weak if they may be easily guessed by an attacker. iOS, iPadOS,
and macOS detect common patterns used to create memorable passwords, such as using
words found in a dictionary, common character substitutions (such as using “p4ssw0rd”
instead of “password”), patterns found on a keyboard (such as “q12we34r” from a
QWERTY keyboard), or repeated sequences (such as “123123”). These patterns are often
used to create passwords that satisfy minimum password requirements for services, but
are also commonly used by attackers attempting to obtain a password using brute force.
Because many services specifically require a four- or six-digit PIN code, these short
passcodes are evaluated with different rules. PIN codes are considered weak if they are
one of the most common PIN codes, if they are an increasing or decreasing sequence such
as “1234” or “8765,” or if they follow a repetition pattern, such as “123123” or “123321.”
Passwords are marked leaked if the Password Monitoring feature can claim they have been
present in a data leak. For more information, see Password Monitoring.
Weak, reused, and leaked passwords are either indicated in the list of passwords (macOS)
or present in the dedicated Security Recommendations interface (iOS and iPadOS). If the
user logs in to a website in Safari using a previously saved password that’s very weak or
that’s been compromised by a data leak, they’re shown an alert strongly encouraging them
to upgrade to an automatic strong password.

Upgrading account authentication security with Account Authentication
Modification Extensions
Apps that implement an Account Authentication Modification Extension can provide easy,
tap-of-a-button upgrades for password-based accounts—namely, they can switch to using
Sign in with Apple or an automatic strong password. This extension point is available in iOS
and iPadOS.
If an app has implemented the extension point and is installed on device, users see
extension upgrade options when viewing Security Recommendations for credentials
associated with the app in the iCloud Keychain password manager in Settings. The
upgrades are also offered when users sign in to the app with the at-risk credential. Apps
have the ability to tell the system not to prompt users with upgrade options after signing in.
Using new AuthenticationServices API, apps can also invoke their extensions and perform
upgrades themselves, ideally from an account settings or account management screen in
the app.
Apps can choose to support strong password upgrades, Sign in with Apple upgrades, or
both. In a strong password upgrade, the system generates an automatic strong password
for the user. If necessary, the app can provide custom password rules to follow when
generating the new password. When a user switches an account from using a password to
using Sign in with Apple, the system provides a new Sign in with Apple credential to the
extension to associate the account with. The user’s Apple ID email isn’t provided as part
of the credential. After a successful Sign in with Apple upgrade, the system deletes the
previously used password credential from the user’s keychain if it’s saved there.
Account Authentication Modification Extensions have the opportunity to perform additional

user authentication before performing an upgrade. For upgrades initiated within the
password manager or after signing in to an app, the extension provides the user name and
password for the account to upgrade. For in-app upgrades, only the user name is provided.
If the extension requires further user authentication, it can request to show a custom user
interface before moving on with the upgrade. The intended use case for showing this
user interface is to have the user enter a second factor of authentication to authorize the
upgrade.
Password Monitoring
Password Monitoring is a feature that matches passwords stored in your Password AutoFill
keychain against a continuously updated and curated list of passwords known to have
been exposed in leaks from different online organizations. If the feature is turned on, the
monitoring protocol continuously matches your Password AutoFill keychain passwords
against the curated list.
How monitoring works

The user’s device continuously performs round-robin checks on a user’s passwords,
querying on an interval that’s independent of the user’s passwords, ensuring that
verification states remain up to date with the current curated list of leaked passwords. To
prevent leakage of information related to how many unique passwords a user has, requests
are batched and performed in parallel. A fixed number of passwords are verified in parallel
on each check, and should the user have fewer than this number, random passwords are
generated and added to the queries to make up the difference.

How matching is done
Matching is done in a two-part process. The most commonly leaked passwords are
contained within a local list on the user’s device. If the user’s password occurs on this
list, the user is immediately notified without any external interaction. This ensures that
no information is leaked about the passwords a user has that are most at risk due to a
password breach.
If the password isn’t contained on the most frequent list, then it is matched against less
frequently leaked passwords.
Comparing users’ passwords against a curated list

To verify whether a password not present in the local list is a match involves some
interaction with Apple servers. To ensure that legitimate users’ passwords aren’t sent
to Apple, a form of cryptographic private set intersection is deployed that compares the
users’ passwords against a large set of leaked passwords. This ensures that for passwords
less at risk of breach, little information is shared with Apple. For a user’s password, this
information is limited to a 15-bit prefix of a cryptographic hash. The removal of the most
frequently leaked passwords from this interactive process, using the local list of most
commonly leaked passwords, reduces the delta in relative frequency of passwords in the
web services buckets, making it impractical to infer user passwords from these lookups.
The underlying protocol partitions the list of curated passwords, which contained
approximately 1.5 billion passwords at the time of this writing, into 215 different buckets.
The bucket a password belongs to is based on the first 15 bits of the SHA256 hash value of
the password. Additionally, each leaked password, pw, is associated with an elliptic curve
point on the NIST P256 curve: Ppw = ⍺·HSWU(pw), where ⍺ is a secret random key known
only to Apple and HSWU is a random oracle function that maps passwords to curve points
based on the Shallue-van de Woestijne-Ulas method. This transformation is designed
to computationally hide the values of passwords and prevents revealing newly leaked
passwords through Password Monitoring.
To compute the private set intersection, the user’s device determines the bucket the user’s
password belongs to using λ, the 15-bit prefix of SHA256(upw), where upw is one of the
user’s passwords. The device generates their own random constant, β, and sends the point
Pc = β·HSWU(upw) to the server, along with a request for the bucket corresponding to λ.
Here β hides information about the user’s password and limits to λ the information exposed
from the password to Apple. Finally, the server takes the point sent by the user’s device,
computes ⍺Pc = ⍺β·HSWU(upw), and returns it, along with the appropriate bucket of points—
Bλ = { Ppw | SHA256(pw) begins with prefix λ}—to the device.
The returned information allows the device to compute B’λ = {β·Ppw | Ppw ∈ Bλ}, and
ascertains that the user’s password has been leaked if ⍺Pc ∈ B'λ.
Sending passwords to other users or Apple devices
Saving credentials to another device with AirDrop

When iCloud is enabled, users can AirDrop a saved credential—including the websites it’s
saved for, its user name, and its password—to another device. Sending credentials with
AirDrop always operates in Contacts Only mode, regardless of the user’s settings. On
the receiving device, after user consent, the credential are stored in the user’s Password
AutoFill Keychain.

Filling in credentials in apps on Apple TV
Password AutoFill is available to fill credentials in apps on Apple TV. When the user focuses
on a user name or password text field in tvOS, Apple TV begins advertising a request for
Password AutoFill over Bluetooth Low Energy (BLE).
Any nearby iPhone, iPad, or iPod touch displays a prompt inviting the user to share a
credential with Apple TV. Here’s how the encryption method is established:
• If the device and Apple TV uses the same iCloud account, encryption between the
devices happens automatically.
• If the device is signed in to an iCloud account other than the one used by Apple TV, the
user is prompted to establish an encrypted connection through use of a PIN code. To
receive this prompt, iPhone must be unlocked and in close proximity to the Siri Remote
paired to that Apple TV.
After the encrypted connection is made using BLE link encryption, the credential is sent to
Apple TV and is automatically filled in to the relevant text fields on the app.
Credential provider extensions

In iOS, iPadOS, and macOS, users can designate a participating third-party app as a
credential provider for Password AutoFill in Passwords settings (iOS and iPadOS) or in
Extensions settings in System Preferences (macOS). This mechanism is built on app
extensions. The credential provider extension must provide a view for choosing credentials,
and the extension can optionally provide metadata about saved credentials so they can be
offered directly on the QuickType bar (iOS and iPadOS) or in an autocomplete suggestion
(macOS). The metadata includes the website of the credential and the associated user
name but not its password. iOS, iPadOS, and macOS communicate with the extension to
get the password when the user chooses to fill a credential into an app or a website in
Safari. Credential metadata is stored inside the credential provider app’s container and is
automatically removed when an app is uninstalled.
iCloud Keychain
iCloud Keychain security overview

iCloud allows users to securely sync their passwords between iOS and iPadOS devices and
Mac computers without exposing that information to Apple. In addition to strong privacy
and security, other goals that heavily influenced the design and architecture of iCloud
Keychain were ease of use and the ability to recover a keychain. iCloud Keychain consists
of two services: keychain syncing and keychain recovery.
Apple designed iCloud Keychain and keychain recovery so that a user’s passwords are still
protected under the following conditions:
• A user’s iCloud account is compromised.
• iCloud is compromised by an external attacker or employee.
• A third party accesses user accounts.

Password manager integration with iCloud Keychain
iOS, iPadOS, and macOS can automatically generate cryptographically strong random
strings to use as account passwords in Safari. iOS and iPadOS can also generate strong
passwords for apps. Generated passwords are stored in the keychain and synced to other
devices. Keychain items are transferred from device to device, traveling through Apple
servers, but are encrypted in such a way that Apple and other devices can’t read their
contents.
Secure keychain syncing

When a user enables iCloud Keychain for the first time, the device establishes a circle of
trust and creates a syncing identity for itself. The syncing identity consists of a private key
and a public key. The public key of the syncing identity is put in the circle, and the circle is
signed twice: first by the private key of the syncing identity, then again with an asymmetric
elliptical key (using P-256) derived from the user’s iCloud account password. Also stored
with the circle are the parameters (random salt and iterations) used to create the key that
is based on the user’s iCloud password.
Syncing circle is placed in user’s iCloud

The signed syncing circle is placed in the user’s iCloud key-value storage area. It can’t be
read without knowing the user’s iCloud password and can’t be modified validly without
having the private key of the syncing identity of its member.
When the user turns on iCloud Keychain on another device, iCloud Keychain notices that
the user has a previously established syncing circle in iCloud that it isn’t a member of. The
device creates its syncing identity key pair, then creates an application ticket to request
membership in the circle. The ticket consists of the device’s public key of its syncing
identity, and the user is asked to authenticate with their iCloud password. The elliptical
key-generation parameters are retrieved from iCloud and generate a key that’s used to
sign the application ticket. Finally, the application ticket is placed in iCloud.
How a user’s other devices are added to the syncing circle

When the first device sees that an application ticket has arrived, it asks for the user to
acknowledge that a new device is asking to join the syncing circle. The user enters their
iCloud password, and the application ticket is verified as signed by a matching private key.
Now, the users who generated the request to join the circle can join it.
Upon the user’s approval to add the new device to the circle, the first device adds the
public key of the new member to the syncing circle and signs it again with both its syncing
identity and the key derived from the user’s iCloud password. The new syncing circle is
placed in iCloud, where it’s similarly signed by the new member of the circle.
There are now two members of the signing circle, and each member has the public key of
its peer. They now begin to exchange individual keychain items through iCloud key-value
storage or store them in CloudKit, whichever is most appropriate for the situation. If both
circle members have the same item, the one with the most recent modification date is
synced. If the other member has the item and the modification dates are identical, items
are skipped. Each item that’s synced is encrypted so that it can be decrypted only by a
device within the user’s circle of trust; it can’t be decrypted by any other devices or by
Apple.

This process is repeated as new devices join the syncing circle. For example, when a third
device joins, the confirmation appears on both of the other user’s devices. The user can
approve the new member from either of those devices. As new peers are added, each peer
syncs with the new one to ensure that all members have the same keychain items.
Only certain items are synced

However, the entire keychain isn’t synced. Some items are device specific, such as VPN
identities, and shouldn’t leave the device. Only items with the kSecAttrSynchronizable
attribute are synced. Apple has set this attribute for Safari user data (including user
names, passwords, and credit card numbers) as well as for Wi-Fi passwords and HomeKit
encryption keys.
Additionally, by default, keychain items added by third-party apps don’t sync. Developers
must set the kSecAttrSynchronizable attribute when adding items to the keychain.
Secure iCloud Keychain recovery

iCloud Keychain recovery provides a way for users to optionally escrow their keychain with
Apple without allowing Apple to read the passwords and other data it contains. Even if the
user has only a single device, keychain recovery provides a safety net against data loss.
This is particularly important when Safari is used to generate random, strong passwords for
web accounts, because the only record of those passwords is in the keychain.
A cornerstone of keychain recovery is secondary authentication and a secure escrow

service, created by Apple specifically to support this feature. The user’s keychain is
encrypted using a strong passcode, and the escrow service provides a copy of the
keychain only if a strict set of conditions are met.
Use of secondary authentication

There are several ways to establish a strong passcode:
• If two-factor authentication is enabled for the user’s account, the device passcode is
used to recover an escrowed keychain.
• If two-factor authentication isn’t set up, the user is asked to create an iCloud security
code by providing a six-digit passcode. Alternatively, without two-factor authentication,
users can specify their own, longer code, or they can let their devices create a
cryptographically random code that they can record and keep on their own.
Keychain escrow process

After the passcode is established, the keychain is escrowed with Apple. The iOS, iPadOS,
or macOS device first exports a copy of the user’s keychain and then encrypts it wrapped
with keys in an asymmetric keybag and places it in the user’s iCloud key-value storage
area. The keybag is wrapped with the user’s iCloud security code and with the public
key of the hardware security module (HSM) cluster that stores the escrow record. This
becomes the user’s iCloud escrow record. For HSA2 accounts, the keychain is also stored
in CloudKit and wrapped to intermediate keys that are recoverable only with the contents
of the iCloud escrow record, thereby providing the same level of protection.
Note: If the user decides to accept a cryptographically random security code instead of
specifying their own or using a four-digit value, no escrow record is necessary. Instead, the
iCloud security code is used to wrap the random key directly.

Besides establishing a security code, users must register a phone number. This provides
a secondary level of authentication during keychain recovery. The user receives an SMS
message that must be replied to for the recovery to proceed.
Escrow security for iCloud Keychain

iCloud provides a secure infrastructure for keychain escrow to ensure that only authorized
users and devices can perform a recovery. Topographically positioned behind iCloud are
clusters of hardware security modules (HSMs) that guard the escrow records. As described
previously, each has a key that is used to encrypt the escrow records under their watch.
To recover a keychain, users must authenticate with their iCloud account and password and
respond to an SMS sent to their registered phone number. After this is done, users must
enter their iCloud security code. The HSM cluster verifies that a user knows their iCloud
security code using the Secure Remote Password (SRP) protocol; the code itself isn’t sent
to Apple. Each member of the cluster independently verifies that the user hasn’t exceeded
the maximum number of attempts allowed to retrieve their record, as discussed below. If a
majority agree, the cluster unwraps the escrow record and sends it to the user’s device.
Next, the device uses the iCloud security code to unwrap the random keys used to encrypt
the user’s keychain. With that key, the keychain—retrieved from iCloud key-value storage
and CloudKit—is decrypted and restored onto the device. iOS, iPadOS, and macOS allow
only 10 attempts to authenticate and retrieve an escrow record. After several failed
attempts, the record is locked and the user must call Apple Support to be granted more
attempts. After the 10th failed attempt, the HSM cluster destroys the escrow record and
the keychain is lost forever. This provides protection against a brute-force attempt to
retrieve the record, at the expense of sacrificing the keychain data in response.
These policies are coded in the HSM firmware. The administrative access cards that permit
the firmware to be changed have been destroyed. Any attempt to alter the firmware or
access the private key causes the HSM cluster to delete the private key. Should this occur,
the owner of each keychain protected by the cluster receives a message informing them
that their escrow record has been lost. They can then choose to reenroll.
Apple Pay
Apple Pay security overview
With Apple Pay, users can use supported iPhone, iPad, Mac, and Apple Watch devices to
pay in an easy, secure, and private way in stores, apps, and on the web in Safari. Users can
also add Apple Pay–enabled transit and student ID cards to Apple Wallet. It’s simple for
users, and it’s built with integrated security in both hardware and software.
Apple Pay is also designed to protect the user’s personal information. Apple Pay doesn’t
collect any transaction information that can be tied back to the user. Payment transactions
are between the user, the merchant, and the card issuer.

Apple Pay component security
Secure Element
The Secure Element is an industry-standard, certified chip running the Java Card platform,
which is compliant with financial industry requirements for electronic payments. The
Secure Element IC and the Java Card platform are certified in accordance with the EMVCo
Security Evaluation process. After the successful completion of the security evaluation,
EMVCo issues unique IC and platform certificates.
The Secure Element IC has been certified based on the Common Criteria standard.
NFC controller
The NFC controller handles Near Field Communication protocols and routes communication
between the Application Processor and the Secure Element, and between the Secure
Element and the point-of-sale terminal.
Apple Wallet
Apple Wallet is used to add and manage credit, debit, and store cards and to make
payments with Apple Pay. Users can view their cards and may be able to view additional
information provided by their card issuer, such as their card issuer’s privacy policy, recent
transactions, and more in Apple Wallet. Users can also add cards to Apple Pay in:
• Setup Assistant and Settings for iOS and iPadOS
• The Watch app for Apple Watch
• Wallet & Apple Pay in System Preferences for Mac computers with Touch ID
In addition, Apple Wallet allows users to add and manage transit cards, rewards cards,
boarding passes, tickets, gift cards, student ID cards, and more.
Secure Enclave
On iPhone, iPad, Apple Watch, and Mac computers with Touch ID, the Secure Enclave
manages the authentication process and enables a payment transaction to proceed.
On Apple Watch, the device must be unlocked, and the user must double-click the side
button. The double-click is detected and passed directly to the Secure Element or Secure
Enclave, where available, without going through the Application Processor.
Apple Pay servers

The Apple Pay servers manage the setup and provisioning of credit, debit, transit, and
student ID cards in the Wallet app. The servers also manage the Device Account Numbers
stored in the Secure Element. They communicate both with the device and with the
payment network or card issuer servers. The Apple Pay servers are also responsible for
reencrypting payment credentials for payments within apps or on the web.

How Apple Pay uses the Secure Element and NFC controller
Secure Element
The Secure Element hosts a specially designed applet to manage Apple Pay. It also
includes applets certified by payment networks or card issuers. Credit, debit, or prepaid
card data is sent from the payment network or card issuer encrypted to these applets
using keys that are known only to the payment network or card issuer and the applets’
security domain. This data is stored within these applets and protected using the Secure
Element’s security features. During a transaction, the terminal communicates directly
with the Secure Element through the near-field-communication (NFC) controller over a
dedicated hardware bus.
NFC controller
As the gateway to the Secure Element, the NFC controller ensures that all contactless
payment transactions are conducted using a point-of-sale terminal that is in close
proximity with the device. Only payment requests arriving from an in-field terminal are
marked by the NFC controller as contactless transactions.
After a credit, debit, or prepaid card (including store cards) payment is authorized by
the cardholder using Touch ID, Face ID, or a passcode, or on an unlocked Apple Watch
by double-clicking the side button, contactless responses prepared by the payment
applets within the Secure Element are exclusively routed by the controller to the NFC
field. Consequently, payment authorization details for contactless payment transactions
are contained to the local NFC field and are never exposed to the Application Processor.
In contrast, payment authorization details for payments within apps and on the web are
routed to the Application Processor, but only after encryption by the Secure Element to the
Apple Pay server.
Credit, debit, and prepaid cards
Card provisioning security overview

When a user adds a credit, debit, or prepaid card (including store cards) to Apple Wallet,
Apple securely sends the card information, along with other information about user’s
account and device, to the card issuer or card issuer’s authorized service provider. Using
this information, the card issuer determines whether to approve adding the card to
Apple Wallet.
As part of the card provisioning process, Apple Pay uses three server-side calls to send
and receive communication with the card issuer or network: Required Fields, Check Card,
and Link and Provision. The card issuer or network uses these calls to verify, approve, and
add cards to Apple Wallet. These client-server sessions use TLS 1.2 to transfer the data.
Full card numbers aren’t stored on the device or on Apple Pay servers. Instead, a unique
Device Account Number is created, encrypted, and then stored in the Secure Element. This
unique Device Account Number is encrypted in such a way that Apple can’t access it. The
Device Account Number is unique and different from most credit or debit card numbers;
the card issuer or payment network can prevent its use on a magnetic stripe card, over the
phone, or on websites. The Device Account Number in the Secure Element is never stored
on Apple Pay servers or backed up to iCloud, and it is isolated from iOS, iPadOS, and
watchOS devices and from Mac computers with Touch ID.

Cards for use with Apple Watch are provisioned for Apple Pay using the Apple Watch app
on iPhone, or within a card issuer’s iPhone app. Adding a card to Apple Watch requires
that the watch be within Bluetooth communications range. Cards are specifically enrolled
for use with Apple Watch and have their own Device Account Numbers, which are stored
within the Secure Element on the Apple Watch.
When credit, debit, or prepaid cards (including store cards) are added, they appear in a list
of cards during Setup Assistant on devices that are signed in to the same iCloud account.
These cards remain in this list for as long as they are active on at least one device. Cards
are removed from this list after they have been removed from all devices for 7 days. This
feature requires two-factor authentication to be enabled on the respective iCloud account.
Adding credit or debit cards to Apple Pay
Adding credit or debit cards manually

To add a card manually, the name, card number, expiration date, and CVV are used
to facilitate the provisioning process. From within Settings, the Wallet app, or the
Apple Watch app, users can enter that information either by typing or by using the device’s
camera. When the camera captures the card information, Apple attempts to populate the
name, card number, and expiration date. The photo is never saved to the device or stored
in the photo library. After all the fields are filled in, the Check Card process verifies the
fields other than the CVV. They are then encrypted and sent to the Apple Pay server.
If a terms and conditions ID is returned with the Check Card process, Apple downloads
and displays the terms and conditions of the card issuer to the user. If the user accepts
the terms and conditions, Apple sends the ID of the terms that were accepted as well as
the CVV to the Link and Provision process. Additionally, as part of the Link and Provision
process, Apple shares information from the device with the card issuer or network, like
information about the user’s iTunes and App Store account activity (for example, whether
the user has a long history of transactions within iTunes), information about the user’s
device (for example, phone number, name, and model of the user’s device plus any
companion Apple device necessary to set up Apple Pay), and the user’s approximate
location at the time the user adds their card (if the user has Location Services enabled).
Using this information, the card issuer determines whether to approve adding the card to
Apple Pay.
As the result of the Link and Provision process, two things occur:
• The device begins to download the Wallet pass file representing the credit or debit card.
• The device begins to bind the card to the Secure Element.
The pass file contains URLs to download card art, metadata about the card such as contact
information, the related issuer’s app, and supported features. It also contains the pass
state, which includes information such as whether the personalizing of the Secure Element
has completed, whether the card is currently suspended by the card issuer, or whether
additional verification is required before the card can make payments with Apple Pay.

Adding credit or debit cards from an iTunes Store account
For a credit or debit card on file with iTunes, the user may be required to reenter their
Apple ID password. The card number is retrieved from iTunes, and the Check Card process
is initiated. If the card is eligible for Apple Pay, the device downloads and displays terms
and conditions, then send along the term’s ID and the card security code to the Link and
Provision process. Additional verification may occur for iTunes account cards on file.
Adding credit or debit cards from a card issuer’s app

When the app is registered for use with Apple Pay, keys are established for the app and for
the card issuer’s server. These keys are used to encrypt the card information that’s sent to
the card issuer, which prevents the information from being read by the Apple device. The
provisioning flow is similar to that used for manually added cards, described previously,
except one-time passwords are used in lieu of the CVV.
Adding additional verification

A card issuer can decide whether a credit or debit card requires additional verification.
Depending on what’s offered by the card issuer, the user may be able to choose between
different options for additional verification, such as a text message, email, customer
service call, or a method in an approved third-party app to complete the verification. For
text messages or email, the user selects from contact information the issuer has on file.
A code is sent, which must be entered into the Wallet app, Settings, or the Apple Watch
app. For customer service or verification using an app, the issuer performs their own
communication process.
Payment authorization with Apple Pay

For devices having a Secure Enclave, a payment can be made only after it receives
authorization from the Secure Enclave. On iPhone or iPad, this involves confirming that the
user has authenticated with Touch ID, Face ID, or the device passcode. Touch ID or Face ID,
if available, is the default method, but the passcode can be used at any time. A passcode
is automatically offered after three unsuccessful attempts to match a fingerprint, or two
unsuccessful attempts to match a face; after five unsuccessful attempts, the passcode
is required. A passcode is also required when Touch ID or Face ID isn’t configured or isn’t
enabled for Apple Pay. For a payment to be made on Apple Watch, the device must be
unlocked with passcode and the side button must be double-clicked.
Using a shared pairing key

Communication between the Secure Enclave and the Secure Element takes place over a
serial interface, with the Secure Element connected to the NFC controller, which in turn
is connected to the Application Processor. Though not directly connected, the Secure
Enclave and Secure Element can communicate securely using a shared pairing key that’s
provisioned during the manufacturing process. The encryption and authentication of the
communication are based on AES, with cryptographic nonces used by both sides to protect
against replay attacks. The pairing key is generated inside the Secure Enclave from its UID
key and the Secure Element unique identifier. The pairing key is then securely transferred
from the Secure Enclave to a hardware security module (HSM) in the factory, which has the
key material required to then inject the pairing key into the Secure Element.

Authorizing a secure transaction
When the user authorizes a transaction, which includes a physical gesture communicated
directly to the Secure Enclave, the Secure Enclave then sends signed data about the
type of authentication and details about the type of transaction (contactless or within
apps) to the Secure Element, tied to an Authorization Random (AR) value. The AR value
is generated in the Secure Enclave when a user first provisions a credit card and persists
while Apple Pay is enabled, protected by the Secure Enclave encryption and anti-rollback
mechanism. It’s securely delivered to the Secure Element by leveraging the pairing key.
On receipt of a new AR value, the Secure Element marks any previously added cards as
deleted.
Using a payment cryptogram for dynamic security

Payment transactions originating from the payment applets include a payment cryptogram
along with a Device Account Number. This cryptogram, a one-time code, is computed
using a transaction counter and a key. The transaction counter is incremented for each
new transaction. The key is provisioned in the payment applet during personalization and is
known by the payment network and/or the card issuer. Depending on the payment scheme,
other data may also be used in the calculation, including:
• A Terminal Unpredictable Number, for near-field-communication (NFC) transactions
• An Apple Pay server nonce, for transactions within apps
These security codes are provided to the payment network and to the card issuer, which
allows the issuer to verify each transaction. The length of these security codes may vary
based on the type of transaction.
Paying with cards using Apple Pay
Paying with cards in stores

If iPhone or Apple Watch is on and detects an NFC field, it presents the user with the
requested card (if automatic selection is turned on for that card) or the default card, which
is managed in Settings. The user can also go to the Wallet app and choose a card, or when
the device is locked, can:
• Double-click the Home button on devices with Touch ID
• Double-click the side button on devices with Face ID
Next, before information is transmitted, the user must authenticate using Touch ID,
Face ID, or their passcode. When Apple Watch is unlocked, double-clicking the side
button activates the default card for payment. No payment information is sent without user
authentication.
After the user authenticates, the Device Account Number and a transaction-specific
dynamic security code are used when processing the payment. Neither Apple nor a user’s
device sends the full actual credit or debit card numbers to merchants. Apple may receive
anonymous transaction information such as the approximate time and location of the
transaction, which helps improve Apple Pay and other Apple products and services.

Paying with cards within apps
Apple Pay can also be used to make payments on iPhone, iPad, Mac, and Apple Watch apps.
When users pay within apps using Apple Pay, Apple receives the encrypted transaction
information. Before that information is sent to the developer or merchant, Apple reencrypts
the transaction with a developer-specific key. Apple Pay retains anonymous transaction
information, such as approximate purchase amount. This information can’t be tied to the
user and never includes what the user is buying.
When an app initiates an Apple Pay payment transaction, the Apple Pay servers receive the
encrypted transaction from the device prior to the merchant receiving it. The Apple Pay
servers then reencrypt the transaction with a merchant-specific key before relaying it to
the merchant.
When an app requests a payment, it calls an API to determine whether the device supports
Apple Pay and whether the user has credit or debit cards that can make payments on a
payment network accepted by the merchant. The app requests any pieces of information it
needs to process and fulfill the transaction, such as the billing and shipping address, and
contact information. The app then asks iOS, iPadOS, or watchOS to present the Apple Pay
sheet, which requests information for the app as well as other necessary information, such
as the card to use.
At this time, the app is presented with city, state, and zip code information to calculate
the final shipping cost. The full set of requested information isn’t provided to the app until
the user authorizes the payment with Touch ID, Face ID, or the device passcode. After the
payment is authorized, the information presented in the Apple Pay sheet is transferred to
the merchant.
App payment authorization

When the user authorizes the payment, a call is made to the Apple Pay servers to obtain
a cryptographic nonce, which is similar to the value returned by the NFC terminal used
for in-store transactions. The nonce, along with other transaction data, is passed to the
Secure Element to generate a payment credential that’s encrypted with an Apple key.
When the encrypted payment credential comes out of the Secure Element, it’s passed to
the Apple Pay servers, which decrypt the credential, verify the nonce in the credential
against the nonce originally sent by the Apple Pay servers, and reencrypt the payment
credential with the merchant key associated with the Merchant ID. The payment is then
returned to the device, which hands it back to the app through the API. The app then
passes it along to the merchant system for processing. The merchant can then decrypt the
payment credential with its private key for processing. This, together with the signature
from Apple’s servers, allows the merchant to verify that the transaction was intended for
this particular merchant.
The APIs require an entitlement that specifies the supported Merchant IDs. An app can
also include additional data (such as an order number or customer identity) to send to the
Secure Element to be signed, ensuring that the transaction can’t be diverted to a different
customer. This is accomplished by the app developer, who can specify applicationData on
the PKPaymentRequest. A hash of this data is included in the encrypted payment data. The
merchant is then responsible for verifying that their applicationData hash matches what’s
included in the payment data.

Paying with cards at websites
Apple Pay can be used to make payments at websites on iPhone, iPad, Apple Watch, and
Mac computers with Touch ID. Apple Pay transactions can also start on a Mac and be
completed on an Apple Pay–enabled iPhone or Apple Watch using the same iCloud account.
Apple Pay on the web requires that all participating websites register with Apple. After
the domain is registered, domain name validation is performed only after Apple issues a
TLS client certificate. Websites supporting Apple Pay are required to serve their content
over HTTPS. For each payment transaction, websites need to obtain a secure and unique
merchant session with an Apple server using the Apple-issued TLS client certificate.
Merchant session data is signed by Apple. After a merchant session signature is verified, a
website may query whether the user has an Apple Pay–capable device and whether they
have a credit, debit, or prepaid card activated on the device. No other details are shared.
If the user doesn’t want to share this information, they can disable Apple Pay queries in
Safari privacy settings on iPhone, iPad, and Mac devices.
After a merchant session is validated, all security and privacy measures are the same as
when a user pays within an app.
If the user is transmitting payment-related information from a Mac to an iPhone or

Apple Watch, Apple Pay Handoff uses the end-to-end encrypted Apple Identity Service
(IDS) protocol to transmit payment-related information between the user’s Mac and
the authorizing device. IDS uses the user’s device keys to perform encryption so no
other device can decrypt this information, and the keys aren’t available to Apple. Device
discovery for Apple Pay Handoff contains the type and unique identifier of the user’s
credit cards along with some metadata. The device-specific account number of the user’s
card isn’t shared, and it continues to remain stored securely on the user’s iPhone or
Apple Watch. Apple also securely transfers the user’s recently used contact, shipping, and
billing addresses over iCloud Keychain.
After the user authorizes payment using Touch ID, Face ID, a passcode, or double-clicking
the side button on Apple Watch, a payment token uniquely encrypted to each website’s
merchant certificate is securely transmitted from the user’s iPhone or Apple Watch to their
Mac and then delivered to the merchant’s website.
Only devices in proximity to each other may request and complete payment. Proximity is
determined through Bluetooth Low Energy (BLE) advertisements.
Contactless passes in Apple Pay

To transmit data from supported passes to compatible NFC terminals, Apple uses the
Apple Value Added Services (Apple VAS) protocol. The VAS protocol can be implemented
on contactless terminals and uses NFC to communicate with supported Apple devices. The
VAS protocol works over a short distance and can be used to present contactless passes
independently or as part of an Apple Pay transaction.
When the device is held near the NFC terminal, the terminal initiates receiving the pass
information by sending a request for a pass. If the user has a pass with the pass provider’s
identifier, the user is asked to authorize its use using Touch ID, Face ID, or a passcode. The
pass information, a timestamp, and a single-use random ECDH P-256 key are used with
the pass provider’s public key to derive an encryption key for the pass data, which is sent
to the terminal.

From iOS 12 to and including iOS 13, users may manually select a pass before presenting it
to the merchant’s NFC terminal. In iOS 13.1 or later, pass providers can configure manually
selected passes to either require user authentication or be used without authentication.
Rendering cards unusable with Apple Pay

Credit, debit, and prepaid cards added to the Secure Element can be used only if
the Secure Element is presented with authorization using the same pairing key and
Authorization Random (AR) value from when the card was added. On receipt of a new AR
value, the Secure Element marks any previously added cards as deleted. This allows the
operating system to instruct the Secure Enclave to render cards unusable by marking its
copy of the AR as invalid under the following scenarios:
Method Device
When the passcode is disabled iPhone, iPad, Apple Watch
When the password is disabled Mac
The user signs out of iC loud iPhone, iPad, Mac, Apple Watch
The user selects Erase All Content and Settings iPhone, iPad, Apple Watch
The device is restored from Recovery mode iPhone, iPad, Mac, Apple Watch
Unpairing Apple Watch
Suspending, removing, and erasing cards

Users can suspend Apple Pay on iPhone, iPad, and Apple Watch by placing their devices
in Lost Mode using Find My. Users also have the ability to remove and erase their cards
from Apple Pay using Find My, iCloud.com, or directly on their devices using the Wallet
app. On Apple Watch, cards can be removed using iCloud settings, the Apple Watch app on
iPhone, or directly on the watch. The ability to make payments using cards on the device is
suspended or removed from Apple Pay by the card issuer or respective payment network,
even if the device is offline and not connected to a cellular or Wi-Fi network. Users can
also call their card issuer to suspend or remove cards from Apple Pay.
When a user erases the entire device—using Erase All Content and Settings, using Find My,
or restoring their device—iPhone, iPad, iPod touch, Mac, and Apple Watch instruct the
Secure Element to mark all cards as deleted. This has the effect of immediately changing
the cards to an unusable state until the Apple Pay servers can be contacted to fully erase
the cards from the Secure Element. Independently, the Secure Enclave marks the AR as
invalid so that further payment authorizations for previously enrolled cards aren’t possible.
When the device is online, it attempts to contact the Apple Pay servers to ensure that all
cards in the Secure Element are erased.

Apple Cash security in iOS, iPadOS, and watchOS
Overview
In iOS 11.2 or later, iPadOS 13.1 or later, and watchOS 4.2 or later, Apple Pay can be used
on an iPhone, iPad, or Apple Watch to send, receive, and request money from other users.
When a user receives money, it’s added to an Apple Cash account that can be accessed in
the Wallet app or within Settings > Wallet & Apple Pay across any of the eligible devices
the user has signed in with their Apple ID.
In iOS 14, iPadOS 14, and watchOS 7, the organizer of an iCloud family who has verified
their identity can enable Apple Cash for their family members under the age of 18.
Optionally, the organizer can restrict the money sending capabilities of these users to
family members only or contacts only. If the family member under the age of 18 goes
through an Apple ID account recovery, the organizer of the family must manually reenable
the Apple Cash card for that user. If the family member under the age of 18 is no longer
part of the iCloud family, their Apple Cash balance is automatically transferred to the
organizer’s account.
When the user sets up Apple Cash, the same information as when the user adds a credit or
debit card may be shared with our partner bank Green Dot Bank and with Apple Payments
Inc., a wholly owned subsidiary created to protect the user’s privacy by storing and
processing information separately from the rest of Apple, and in a way that the rest of
Apple doesn’t know. This information is used only for troubleshooting, fraud prevention,
and regulatory purposes.
Using Apple Cash in iMessage

To use person-to-person payments and Apple Cash, a user must be signed in to their
iCloud account on an Apple Cash–compatible device and have two-factor authentication
set up on the iCloud account. Money requests and transfers between users are initiated
from within the Messages app or by asking Siri. When a user attempts to send money,
iMessage displays the Apple Pay sheet. The Apple Cash balance is always used first. If
necessary, additional funds are drawn from a second credit or debit card the user has
added to the Wallet app.
Using Apple Cash in stores, apps, and on the web

The Apple Cash card in the Wallet app can be used with Apple Pay to make payments in
stores, in apps, and on the web. Money in the Apple Cash account can also be transferred
to a bank account. In addition to money being received from another user, money can be
added to the Apple Cash account from a debit or prepaid card in the Wallet app.
Apple Payments Inc. stores, and may use, the user’s transaction data for troubleshooting,
fraud prevention, and regulatory purposes once a transaction is completed. The rest of
Apple doesn’t know who the user sent money to, received money from, or where the user
made a purchase with their Apple Cash card.

When the user sends money with Apple Pay, adds money to an Apple Cash account, or
transfers money to a bank account, a call is made to the Apple Pay servers to obtain a
cryptographic nonce, which is similar to the value returned for Apple Pay within apps. The
nonce, along with other transaction data, is passed to the Secure Element to generate a
payment signature. When the payment signature comes out of the Secure Element, it’s
passed to the Apple Pay servers. The authentication, integrity, and correctness of the
transaction is verified through the payment signature and the nonce by Apple Pay servers.
Money transfer is then initiated, and the user is notified of a completed transaction.
If the transaction involves:
• A debit card for adding money to Apple Cash
• Providing supplemental money if the Apple Cash balance is insufficient
An encrypted payment credential is also produced and sent to Apple Pay servers, similar to
how Apple Pay works within apps and websites.
After the balance of the Apple Cash account exceeds a certain amount or if unusual activity
is detected, the user is prompted to verify their identity. Information provided to verify the
user’s identity—such as social security number or answers to questions (for example, to
confirm a street name the user lived on previously)—is securely transmitted to the Apple
partner and encrypted using their key. Apple can’t decrypt this data. The user is prompted
to verify their identity again if they perform an Apple ID account recovery, before regaining
access to their Apple Cash balance.
Apple Card security
Apple Card application in the Wallet app

In iOS 12.4 or later, macOS 10.14.6 or later, and watchOS 5.3 or later, Apple Card can be
used with Apple Pay to make payments in stores, in apps, and on the web.
To apply for Apple Card, the user must be signed into their iCloud account on an Apple Pay–
compatible iOS or iPadOS device and have two-factor authentication set up on the iCloud
account. When the application is approved, Apple Card is available in the Wallet app or
within Settings > Wallet & Apple Pay across any of the eligible devices the user has signed
in with their Apple ID.
When a user applies for Apple Card, user identity information is securely verified by
Apple’s identity provider partners and then shared with Goldman Sachs Bank USA for the
purposes of identity and credit evaluation.
Information such as the social security number or ID document image provided during the
application is securely transmitted to Apple’s identity provider partners and/or Goldman
Sachs Bank USA encrypted with their respective keys. Apple can’t decrypt this data.
The income information provided during the application, and the bank account information
used for bill payments, are securely transmitted to Goldman Sachs Bank USA encrypted
with their key. The bank account information is saved in the keychain. Apple can’t decrypt
this data.
When adding Apple Card to the Wallet app, the same information as when a user adds a
credit or debit card may be shared with the Apple partner bank Goldman Sachs Bank USA
and with Apple Payments Inc. This information is used only for troubleshooting, fraud
prevention, and regulatory purposes.

A physical card can be ordered from Apple Card in the Wallet app. After the user receives
the physical card, it’s activated using the NFC tag present in the bifold envelope of the
physical card. The tag is unique per card and can’t be used to activate another user’s card.
Alternatively, the card can be manually activated in the Wallet settings. Additionally, the
user can also choose to lock or unlock the physical card at any time from the Wallet app.
Apple Card payments and Apple Wallet pass details

Payments due on the Apple Card account can be made from the Wallet app in iOS with
Apple Cash and a bank account. Bill payments can be scheduled as recurring or as a
one-time payment at a specific date with Apple Cash and a bank account. When a user
makes a payment, a call is made to the Apple Pay servers to obtain a cryptographic nonce
similar to Apple Cash. The nonce, along with the payment setup details, is passed to the
Secure Element to generate a signature. When the payment signature comes out of the
Secure Element, it’s passed to the Apple Pay servers. The authentication, integrity, and
correctness of the payment are verified through the signature and the nonce by Apple Pay
servers, and the order is passed on to Goldman Sachs Bank USA for processing.
Displaying the Apple Card number details in the pass using the Wallet app requires user
authentication with Face ID, Touch ID, or a passcode. It can be replaced by the user in the
card information section and disables the previous one.
Adding transit and student ID cards to Wallet
Transit cards
In many global markets, users can add supported transit cards to the Wallet app on
supported models of iPhone and Apple Watch. Depending on the transit operator, this may
be done by transferring the value and commuter pass from a physical card into its digital
Apple Wallet representation or by provisioning a new transit card into the Wallet app from
the Wallet app or the transit card issuer’s app. After transit cards are added to the Wallet
app, users can ride transit simply by holding their iPhone or Apple Watch near the transit
reader. Some cards can also be used to make payments.
Added transit cards are associated with a user’s iCloud account. If the user adds more than
one card to the Wallet app, Apple or the transit card issuer may be able to link the user’s
personal information and the associated account information between cards. Transit cards
and transactions are protected by a set of hierarchical cryptographic keys.
During the process of transferring the balance from a physical card to the Wallet app,
users are required to enter card specific information. Users may also need to provide
personal information for proof of card possession. When transferring passes from iPhone
to Apple Watch, both devices must be online during transfer.
The balance can be recharged with funds from credit, debit and prepaid cards through
Wallet or from the transit card issuer’s app. To understand the security of reloading the
balance when using Apple Pay, see Paying with cards within apps. To learn how the transit
card is provisioned from within the transit card issuer’s app, see Adding credit or debit
cards from a card issuer’s app.

If provisioning from a physical card is supported, the transit card issuer has the
cryptographic keys needed to authenticate the physical card and verify the user’s entered
data. After the data is verified, the system can create a Device Account Number for the
Secure Element and activate the newly added pass in the Wallet app with the transferred
balance. In some cities, after provisioning from the physical card is complete, the physical
card is disabled.
At the end of either type of provisioning, if the transit card balance is stored on the
device, it’s encrypted and stored to a designated applet in the Secure Element. The transit
operator has the keys to perform cryptographic operations on the card data for balance
transactions.
By default, users benefit from the seamless Express Transit experience that allows them
to pay and ride without requiring Touch ID, Face ID, or a passcode. Information such as
recently visited stations, transaction history, and additional tickets may be accessed by
any nearby contactless card reader with Express Mode enabled. Users can enable the
Touch ID, Face ID, or passcode authorization requirement in the Wallet & Apple Pay settings
by disabling Express Transit.
As with other Apple Pay cards, users can suspend or remove transit cards by:
• Erasing the device remotely with Find My
• Enabling Lost Mode with Find My
• Entering a mobile device management (MDM) remote wipe command
• Removing all cards from their Apple ID account page
• Removing all cards from iCloud.com
• Removing all cards from the Wallet app
• Removing the card in the issuer’s app
Apple Pay servers notify the transit operator to suspend or disable those cards. If a user
removes a transit card from an online device, the balance can be recovered by adding it
back to a device signed in with the same Apple ID. If a device is offline, powered off, or
unusable, recovery may not be possible.
Credit and debit cards

In some cities, transit readers accept EMV (smart) cards to pay for transit rides. When
users present an EMV credit or debit card to those readers, user authentication is required
just as with “Pay with credit and debit cards in the stores.”
In iOS 12.3 or later, some existing EMV credit/debit cards in the Wallet app can be enabled
for Express Transit, which allows the user to pay for a trip at supported transit operators
without requiring Touch ID, Face ID, or a passcode. When a user provisions an EMV credit
or debit card, the first card provisioned to the Wallet app is enabled for Express Transit.
The user can tap the More button on the front of the card in the Wallet app and disable
Express Transit for that card by setting Express Transit Settings to None. The user can
also select a different credit or debit card as their Express Transit card via the Wallet
app. Touch ID, Face ID, or a passcode is required to reenable or select a different card for
Express Transit.
Apple Card and Apple Cash are eligible for Express Transit.

Student ID cards
In iOS 12 or later, students, faculty, and staff at participating campuses can add their
student ID card to the Wallet app on supported models of iPhone and Apple Watch to
access locations and pay wherever their card is accepted.
A user adds their student ID card to the Wallet app through an app provided by the card
issuer or participating school. The technical process by which this occurs is the same as
the one described in Adding credit or debit cards from a card issuer’s app. In addition,
issuing apps must support two-factor authentication on the accounts that guard access to
their student IDs. A card may be set up simultaneously on up to any two supported Apple
devices signed in with the same Apple ID.
When a student ID card is added to the Wallet app, Express Mode is turned on by default.
Student ID cards in Express Mode interact with accepting terminals without Touch ID,
Face ID, passcode authentication, or double-clicking the side button on Apple Watch. The
user can tap the More button on the front of the card in the Wallet app and turn off Express
Mode to disable this feature. Touch ID, Face ID, or a passcode is required to reenable
Express Mode.
Student ID cards can be disabled or removed by:
• Erasing the device remotely with Find My
• Enabling Lost Mode with Find My
• Receiving a mobile device management (MDM) remote wipe command
• Removing all cards from their Apple ID account page
• Removing all cards from iCloud.com
• Removing all cards from the Wallet app
• Removing the card in the issuer’s app
iMessage
iMessage security overview
Apple iMessage is a messaging service for iOS and iPadOS devices, Apple Watch, and Mac
computers. iMessage supports text and attachments such as photos, contacts, locations,
links, and attachments directly on to a message, such as a thumbs up icon. Messages
appear on all of a user’s registered devices so that a conversation can be continued from
any of the user’s devices. iMessage makes extensive use of the Apple Push Notification
service (APNs). Apple doesn’t log the contents of messages or attachments, which are
protected by end-to-end encryption so no one but the sender and receiver can access
them. Apple can’t decrypt the data.
When a user turns on iMessage on a device, the device generates encryption and signing
pairs of keys for use with the service. For encryption, there is an encryption RSA 1280-
bit key as well as an encryption EC 256-bit key on the NIST P-256 curve. For signatures,
Elliptic Curve Digital Signature Algorithm (ECDSA) 256-bit signing keys are used. The
private keys are saved in the device’s keychain and only available after first unlock. The
public keys are sent to Apple Identity Service (IDS), where they are associated with the
user’s phone number or email address, along with the device’s APNs address.

As users enable additional devices for use with iMessage, their encryption and signing
public keys, APNs addresses, and associated phone numbers are added to the directory
service. Users can also add more email addresses, which are verified by sending a
confirmation link. Phone numbers are verified by the carrier network and SIM. With some
networks, this requires using SMS (the user is presented with a confirmation dialog if
the SMS isn’t zero rated). Phone number verification may be required for several system
services in addition to iMessage, such as FaceTime and iCloud. All of the user’s registered
devices display an alert message when a new device, phone number, or email address is
added.
How iMessage sends and receives messages securely

Users start a new iMessage conversation by entering an address or name. If they enter a
phone number or email address, the device contacts the Apple Identity Service (IDS) to
retrieve the public keys and APNs addresses for all of the devices associated with the
addressee. If the user enters a name, the device first uses the user’s Contacts app to
gather the phone numbers and email addresses associated with that name and then gets
the public keys and APNs addresses from IDS.
The user’s outgoing message is individually encrypted for each of the receiver’s devices.
The public encryption keys and signing keys of the receiving devices are retrieved from IDS.
For each receiving device, the sending device generates a random 88-bit value and uses it
as an HMAC-SHA256 key to construct a 40-bit value derived from the sender and receiver
public key and the plaintext. The concatenation of the 88-bit and 40-bit values makes a
128-bit key, which encrypts the message with it using AES in Counter (CTR) Mode. The
40-bit value is used by the receiver side to verify the integrity of the decrypted plaintext.
This per-message AES key is encrypted using RSA-OAEP to the public key of the receiving
device. The combination of the encrypted message text and the encrypted message key
is then hashed with SHA-1, and the hash is signed with the Elliptic Curve Digital Signature
Algorithm (ECDSA) using the sending device’s private signing key. In iOS 13 or later and
iPadOS 13.1 or later, devices may use an Elliptic Curve Integrated Encryption Scheme
(ECIES) encryption instead of RSA encryption.
The resulting messages, one for each receiving device, consist of the encrypted message
text, the encrypted message key, and the sender’s digital signature. They are then
dispatched to the APNs for delivery. Metadata, such as the timestamp and APNs routing
information, isn’t encrypted. Communication with APNs is encrypted using a forward-
secret TLS channel.

APNs can only relay messages up to 4 or 16 KB in size, depending on the iOS or iPadOS
version. If the message text is too long or if an attachment such as a photo is included, the
attachment is encrypted using AES in CTR mode with a randomly generated 256-bit key
and uploaded to iCloud. The AES key for the attachment, its Uniform Resource Identifier
(URI), and an SHA-1 hash of its encrypted form are then sent to the recipient as the
contents of an iMessage, with their confidentiality and integrity protected through normal
iMessage encryption, as shown in the following diagram.
How iMessage sends and receives messages.
For group conversations, this process is repeated for each recipient and their devices.
On the receiving side, each device receives its copy of the message from APNs and, if
necessary, retrieves the attachment from iCloud. The incoming phone number or email
address of the sender is matched to the receiver’s contacts so that a name can be
displayed when possible.
As with all push notifications, the message is deleted from APNs when it’s delivered. Unlike
other APNs notifications, however, iMessage messages are queued for delivery to offline
devices. Messages are stored for up to 30 days.

Secure iMessage name and photo sharing
iMessage Name and Photo Sharing allows a user to share a Name and Photo using
iMessage. The user may select their Me Card information, or customize the name and
include any image they choose. iMessage Name and Photo sharing uses a two-stage
system to distribute the name and photo.
The data is subdivided in fields, each encrypted and authenticated separately as well as
authenticated together with the process below. There are three fields:
• Name
• Photo
• Photo filename
A first step of the data creation is to randomly generate a record 128-bit key on the
device. This record key is then derived with HKDF-HMAC-SHA256 to create three subkeys:
Key 1:Key 2:Key 3 = HKDF(record key, “nicknames”). For each field, a random 96-bit
Initialization Vector (IV) is generated and the data is encrypted using AES-CTR and Key
1. A message authentication code (MAC) is then computed with HMAC-SHA256 using
Key 2 and covering the field name, the field IV, and the field ciphertext. Finally, the set
of individual field MAC values are concatenated and their MAC is computed with HMAC-
SHA256 using Key 3. The 256-bit MAC is stored along side the encrypted data. The first
128 bits of this MAC is used as RecordID.
This encrypted record is then stored in the CloudKit public database under the RecordID.
This record is never mutated, and when the user chooses to change their name and photo,
a new encrypted record is generated each time. When user 1 chooses to share their
name and photo with user 2, they send the record key along with the recordID inside their
iMessage payload, which is encrypted.
When user 2’s device receives this iMessage payload, it notices that the payload contains
a Nickname and Photo recordID and key. User 2’s device then goes out to the public
CloudKit database to retrieve the encrypted name and photo at the record ID and sends it
across using iMessage.
After the message is retrieved, user 2’s device decrypts the payload and verifies the
signature using the recordID itself. If this passes, user 2 is presented with the name and
photo and they can choose to add this to their contacts, or use it for Messages.
Secure Business Chat using the Messages app

Business Chat is a messaging service that enables users to communicate with
businesses using the Messages app. With Business Chat, the user is always in control
of the conversation. They can also delete the conversation and block the business from
messaging them in the future. For privacy, the business doesn’t receive the user’s phone
number, email address, or iCloud account information. Instead, a custom unique identifier
called the Opaque ID is generated by the Apple Identity Service (IDS) and shared with
the business. The Opaque ID is unique to the relationship between the user’s Apple ID
and the business’s Business ID. A user has a different Opaque ID for every business they
contact using Business Chat. The user decides if and when to share personal identifying
information with the business.

Business Chat supports Managed Apple IDs from Apple Business Manager and determines
whether they are enabled for iMessage and FaceTime in Apple School Manager.
Messages sent to the business are encrypted between the user’s device and Apple’s
messaging servers, using the same security and Apple messaging servers as iMessages.
Apple messaging servers decrypt these messages in RAM, and relay them to the business
over an encrypted link using TLS 1.2. Messages are never stored in unencrypted form while
transiting through Apple’s Business Chat service. Businesses’ replies are also sent using
TLS 1.2 to the Apple messaging servers, where they are encrypted using the unique public
keys of each recipient device.
If user devices are online, the message is delivered immediately and isn’t cached on the
Apple messaging servers. If a user’s device isn’t online, the encrypted message is cached
for up to 30 days to enable the user to receive it when the device is back online. As soon
as the device is back online, the message is delivered and deleted from cache. After 30
days, an undelivered cached message expires and is permanently deleted.
The Business Chat service never stores conversation history.
FaceTime security
FaceTime is Apple’s video and audio calling service. Like iMessage, FaceTime calls use
the Apple Push Notification service (APNs) to establish an initial connection to the user’s
registered devices. The audio/video contents of FaceTime calls are protected by end-
to-end encryption, so no one but the sender and receiver can access them. Apple can’t
decrypt the data.
The initial FaceTime connection is made through an Apple server infrastructure that relays
data packets between the users’ registered devices. Using APNs notifications and Session
Traversal Utilities for NAT (STUN) messages over the relayed connection, the devices verify
their identity certificates and establish a shared secret for each session. The shared secret
is used to derive session keys for media channels streamed using the Secure Real-time
Transport Protocol (SRTP). SRTP packets are encrypted using AES256 in Counter Mode
and HMAC-SHA1. Subsequent to the initial connection and security setup, FaceTime uses
STUN and Internet Connectivity Establishment (ICE) to establish a peer-to-peer connection
between devices, if possible.
Group FaceTime extends FaceTime to support up to 33 concurrent participants. As

with classic one-to-one FaceTime, calls are end-to-end encrypted among the invited
participants’ devices. Even though Group FaceTime reuses much of the infrastructure and
design of one-to-one FaceTime, Group FaceTime calls feature a new key-establishment
mechanism built on top of the authenticity provided by Apple Identity Service (IDS). This
protocol provides forward secrecy, meaning the compromise of a user’s device won’t leak
the contents of past calls. Session keys are wrapped using AES-SIV and distributed among
participants using an ECIES construction with ephemeral P-256 ECDH keys.
When a new phone number or email address is added to an ongoing Group FaceTime call,
active devices establish new media keys and never share previously used keys with the
newly invited devices.

Find My
Find My security
Overview
The Find My app combines Find My iPhone and Find My Friends into a single app in iOS,
iPadOS, and macOS. Find My can help users locate a missing device, even an offline Mac.
An online device can simply report its location to the user via iCloud. Find My works offline
by sending out short range Bluetooth signals from the missing device that can be detected
by other Apple devices in use nearby. Those nearby devices then relay the detected
location of the missing device to iCloud so users can locate it in the Find My app—all while
protecting the privacy and security of all the users involved. Find My even works with a
Mac that is offline and asleep.
Using Bluetooth and the hundreds of millions of iOS, iPadOS, and macOS devices in active
use around the world, the user can locate a missing device even if it can’t connect to a
Wi-Fi or cellular network. Any iOS, iPadOS, or macOS device with “offline finding” enabled
in Find My settings can act as a “finder device.” This means the device can detect the
presence of another missing offline device using Bluetooth and then use its network
connection to report an approximate location back to the owner. When a device has offline
finding enabled, it also means that it can be located by other participants in the same way.
This entire interaction is end-to-end encrypted, anonymous, and designed to be battery
and data efficient, so there is minimal impact on battery life cellular data plan usage and
user privacy is protected.
Note: Find My may not be available in all countries or regions.
End-to-end encryption
Find My is built on a foundation of advanced public key cryptography. When offline finding
is enabled in Find My settings, an elliptic curve (EC) P-224 private encryption key pair
noted {d,P} is generated directly on the device where d is the private key and P is the
public key. Additionally, a 256-bit secret SK0 and a counter i is initialized to zero. This
private key pair and the secret are never sent to Apple and are synced only among the
user’s other devices in an end-to-end encrypted manner using iCloud Keychain. The
secret and the counter are used to derive the current symmetric key SKi with the following
recursive construction: SKi = KDF(SKi-1, “update”)
Based on the key SKi, two large integers ui and vi are computed with (ui,vi) = KDF(SKi,
“diversify”). Both the P-224 private key denoted d and corresponding public key referred
to as P are then derived using an affine relation involving the two integers to compute a
short-lived key pair: The derived private key is di, where di = ui * d + vi (modulo the order of
the P-224 curve) and the corresponding public part is Pi and verifies that Pi = ui*P + vi*G.

When a device goes missing and can’t connect to Wi-Fi or cellular—for example, a
MacBook Pro left on a park bench—it begins periodically broadcasting the derived public
key Pi for a limited period of time in a Bluetooth payload. By using P-224, the public
key representation can fit into a single Bluetooth payload. The surrounding devices can
then help in the finding of the offline device by encrypting their location to the public
key. Approximately every 15 minutes, the public key is replaced by a new one using an
incremented value of the counter and the process above so that the user can’t be tracked
by a persistent identifier. The derivation mechanism prevents the various public keys Pi
from being linked to the same device.
Keeping users and devices anonymous

In addition to making sure that location information and other data are fully encrypted,
participants’ identities remain private from each other and from Apple. The traffic sent to
Apple by finder devices contains no authentication information in the contents or headers.
As a result, Apple doesn’t know who the finder is or whose device has been found. Further,
Apple doesn’t log information that would reveal the identity of the finder and retains no
information that would allow anyone to correlate the finder and owner. The device owner
receives only the encrypted location information that’s decrypted and displayed in the
Find My app with no indication as to who found the device.

Using Find My to locate missing Apple devices
Any Apple devices within Bluetooth range that have offline finding enabled can detect
a signal from another Apple device configured to allow Find My and read the current
broadcast key Pi. Using an ECIES construction and the public key Pi from the broadcast,
the finder devices encrypt their current location information and relay it to Apple. The
encrypted location is associated with a server index which is computed as the SHA256
hash of the P-224 public key Pi obtained from the Bluetooth payload. Apple never has the
decryption key, so Apple can’t read the location encrypted by the finder. The owner of the
missing device can reconstruct the index and decrypt the encrypted location.
How Find My locates devices.

When trying to locate the missing device, an expected range of counter values is estimated
for the location search period. With the knowledge of the original private P-224 key d
and secret values SKi in the range of counter values of the search period the owner can
then reconstruct the set of values {di, SHA256(Pi)} for the entire search period. The owner
device used to locate the missing device can then perform queries to the server using the
set of index values SHA256(Pi) and download the encrypted locations from the server. The
Find My app then locally decrypts the encrypted locations with the matching private keys
di and shows an approximate location of the missing device in the app. Location reports
from multiple finder devices are combined by the owner’s app to generate a more precise
location.
How the owner gets the device location from the Find My app.
Locating devices that are offline

If a user has Find My iPhone enabled on their device, offline finding is enabled by default
when they upgrade a device to iOS 13 or later, iPadOS 13.1 or later, and macOS 10.15 or
later. This ensures that every user has the best possible chance to locate their device if it
goes missing. However, if at any time the user prefers not to participate, they can disable
offline finding in Find My settings on their device. When offline finding is disabled, the
device no longer acts as a finder nor is it detectable by other finder devices. However, the
user can still locate the device as long as it can connect to a Wi-Fi or cellular network.
When a missing offline device is located, the user receives a notification and email
message to let them know the device has been found. To view the location of the missing
device, the user opens the Find My app and selects the Devices tab. Rather than showing
the device on a blank map as it would have prior to the device being located, Find My
shows a map location with an approximate address and information on how long ago
the device was detected. If more location reports come in, the current location and time
stamp both update automatically. Although users can’t play a sound on an offline device or
erase it remotely, they can use the location information to retrace their steps or take other
actions to help them recover it.

Continuity
Continuity security overview
Continuity takes advantage of technologies like iCloud, Bluetooth, and Wi-Fi to enable
users to continue an activity from one device to another, make and receive phone calls,
send and receive text messages, and share a cellular internet connection.
Handoff security
Overview
With Handoff, when a user’s iOS, iPadOS, and macOS devices are near each other, the user
can automatically pass whatever they’re working on from one device to the other. Handoff
lets the user switch devices and instantly continue working.
When a user signs in to iCloud on a second Handoff-capable device, the two devices
establish a Bluetooth Low Energy (BLE) 4.2 pairing out-of-band using APNs. The individual
messages are encrypted much like messages in iMessage are. After the devices are paired,
each device generates a symmetric 256-bit AES key that gets stored in the device’s
keychain. This key can encrypt and authenticate the BLE advertisements that communicate
the device’s current activity to other iCloud paired devices using AES256 in GCM mode,
with replay protection measures.
The first time a device receives an advertisement from a new key, it establishes a BLE
connection to the originating device and performs an advertisement encryption key
exchange. This connection is secured using standard BLE 4.2 encryption as well as
encryption of the individual messages, which is similar to how iMessage is encrypted. In
some situations, these messages are sent using APNs instead of BLE. The activity payload
is protected and transferred in the same way as an iMessage.
Handoff between native apps and websites

Handoff allows an iOS, iPadOS, or macOS native app to resume user activity on a webpage
in domains legitimately controlled by the app developer. It also allows the native app user
activity to be resumed in a web browser.
To prevent native apps from claiming to resume websites not controlled by the developer,
the app must demonstrate legitimate control over the web domains it wants to resume.
Control over a website domain is established using the mechanism for shared web
credentials. For details, see App access to saved passwords. The system must validate an
app’s domain name control before the app is permitted to accept user activity Handoff.
The source of a webpage Handoff can be any browser that has adopted the Handoff APIs.
When the user views a webpage, the system advertises the domain name of the webpage
in the encrypted Handoff advertisement bytes. Only the user’s other devices can decrypt
the advertisement bytes.
On a receiving device, the system detects that an installed native app accepts Handoff
from the advertised domain name and displays that native app icon as the Handoff option.
When launched, the native app receives the full URL and the title of the webpage. No other
information is passed from the browser to the native app.

In the opposite direction, a native app may specify a fallback URL when a Handoff
receiving device doesn’t have the same native app installed. In this case, the system
displays the user’s default browser as the Handoff app option (if that browser has adopted
Handoff APIs). When Handoff is requested, the browser is launched and given the fallback
URL provided by the source app. There is no requirement that the fallback URL be limited
to domain names controlled by the native app developer.
Handoff of larger data

In addition to using the basic feature of Handoff, some apps may elect to use APIs that
support sending larger amounts of data over Apple-created peer-to-peer Wi-Fi technology
(much like AirDrop). For example, the Mail app uses these APIs to support Handoff of a
mail draft, which may include large attachments.
When an app uses this facility, the exchange between the two devices starts off just as in
Handoff. However, after receiving the initial payload using Bluetooth Low Energy (BLE), the
receiving device initiates a new connection over Wi-Fi. This connection is encrypted (with
TLS), which exchanges their iCloud identity certificates. The identity in the certificates
is verified against the user’s identity. Further payload data is sent over this encrypted
connection until the transfer is complete.
Universal Clipboard
Universal Clipboard leverages Handoff to securely transfer the content of a user’s
clipboard across devices so they can copy on one device and paste on another. Content is
protected in the same way as other Handoff data and is shared by default with Universal
Clipboard unless the app developer chooses to disallow sharing.
Apps have access to clipboard data regardless of whether the user has pasted the
clipboard into the app. With Universal Clipboard, this data access extends to apps on the
user’s other devices (as established by their iCloud sign-in).
iPhone cellular call relay security

When a user’s Mac, iPad, iPod touch, or HomePod is on the same Wi-Fi network as their
iPhone, it can make and receive phone calls using the cellular connection on iPhone.
Configuration requires the devices to be signed in to both iCloud and FaceTime using the
same Apple ID account.
When an incoming call arrives, all configured devices are notified using the Apple Push
Notification service (APNs), with each notification using the same end-to-end encryption
as iMessage. Devices that are on the same network present the incoming call notification
user interface. When the user answers the call, the audio is seamlessly transmitted from
the user’s iPhone using a secure peer-to-peer connection between the two devices.
When a call is answered on one device, ringing of nearby iCloud-paired devices is

terminated by briefly advertising using Bluetooth Low Energy (BLE). The advertising bytes
are encrypted using the same method as Handoff advertisements.
Outgoing calls are also relayed to iPhone using APNs, and audio is similarly transmitted
over the secure peer-to-peer link between devices. Users can disable phone call relay on a
device by turning off iPhone Cellular Calls in FaceTime settings.

iPhone Text Message Forwarding security
Text Message Forwarding automatically sends SMS text messages received on an iPhone
to a user’s enrolled iPad, iPod touch, or Mac. Each device must be signed in to the
iMessage service using the same Apple ID account. When Text Message Forwarding is
turned on, enrollment is automatic on devices within a user’s circle of trust if two-factor
authentication is enabled. Otherwise, enrollment is verified on each device by entering a
random six-digit numeric code generated by iPhone.
After devices are linked, iPhone encrypts and forwards incoming SMS text messages to
each device, utilizing the methods described in iMessage security overview. Replies are
sent back to iPhone using the same method, and then iPhone sends the reply as a text
message using the carrier’s SMS transmission mechanism. Text Message Forwarding can
be turned on or off in Messages settings.
Instant Hotspot security

Instant Hotspot connects other Apple devices to a personal iOS or iPadOS hotspot. iOS
and iPadOS devices that support Instant Hotspot use Bluetooth Low Energy (BLE) to
discover and communicate to all devices that have signed in to the same individual iCloud
account or accounts used with Family Sharing (in iOS 13 and iPadOS). Compatible Mac
computers with OS X 10.10 or later use the same technology to discover and communicate
with Instant Hotspot iOS and iPadOS devices.
Initially, when a user enters Wi-Fi settings on a device, it emits a BLE advertisement
containing an identifier that all devices signed in to the same iCloud account agree upon.
The identifier is generated from a DSID (Destination Signaling Identifier) that’s tied to the
iCloud account and rotated periodically. When other devices signed in to the same iCloud
account are in close proximity and support Personal Hotspot, they detect the signal and
respond, indicating the availability to use Instant Hotspot.
When a user who isn’t part of Family Sharing chooses an iPhone or iPad for Personal
Hotspot, a request to turn on Personal Hotspot is sent to that device. The request is sent
across a link that is encrypted using BLE encryption, and the request is encrypted in a
fashion similar to iMessage encryption. The device then responds across the same BLE link
using the same per-message encryption with Personal Hotspot connection information.
For users that are part of Family Sharing, Personal Hotspot connection information is
securely shared using a mechanism similar to that used by HomeKit devices to sync
information. Specifically, the connection that shares hotspot information between users is
secured with an ECDH (Curve25519) ephemeral key that is authenticated with the users’
respective device-specific Ed25519 public keys. The public keys used are those that had
previously synced between the members of Family Sharing using IDS when the Family
Share was established.

Car keys security in iOS
Overview
The car keys feature is supported natively in supported iPhone devices and paired
Apple Watch devices. Car keys are represented as passes (created by Apple on behalf of
the automaker) in the Wallet app and support the full Apple Pay card life cycle (iCloud Lost
Mode, Remote Wipe, local pass deletion, and Erase All Content and Settings). In addition
to the standard Apple Pay card management, shared car keys can be deleted from the
owner’s iPhone, Apple Watch, and in the vehicle’s Human Machine Interface (HMI).
Car keys can be used to unlock and lock the vehicle and to start the engine or set the
vehicle into drive mode. The “standard transaction” offers mutual authentication and is
mandatory for engine start. Unlock and lock transactions might use the “fast transaction”
when required for performance reasons.
Keys are created through pairing an iPhone with an owned and supported vehicle. All
keys are created on the embedded Secure Element based on elliptic curve (NIST P-256)
on-board key generation (ECC-OBKG), and the private keys never leave the Secure
Element. Communication between devices and the vehicle use the NFC standard, and key
management uses an Apple to automaker server API with mutually authenticated TLS. After
a key is paired to an iPhone, any Apple Watch paired to that iPhone can also receive a key.
When a key is deleted either in the vehicle or on the device, it can’t be restored. Keys on
lost or stolen devices can be suspended and resumed, but reprovisioning them on a new
device requires a new pairing or sharing.
Owner pairing
The owner must prove possession of the vehicle (the method is dependent on the
automaker) and can start the pairing process in the automaker’s app using an email
link received from the automaker or from the vehicle menu. In all cases, the owner must
present a confidential one-time pairing password to the iPhone, which is used to generate
a secure pairing channel using the SPAKE2+ protocol with the NIST P-256 curve. When
using the app or the email link, the password is automatically transferred to the iPhone
where it must be entered manually when pairing is started from the vehicle.
Key sharing
The owner’s paired iPhone can share keys to eligible family members’ and friends’ iPhone
devices (and their paired Apple Watch devices) by sending a device-specific invitation
using iMessage and the Apple Identity Service (IDS). All sharing commands are exchanged
using the end-to-end encrypted IDS feature. The owner’s paired iPhone keeps the IDS
channel from changing during the sharing process.

Upon acceptance of the invitation, the family member’s or friend’s iPhone creates a digital
key and sends the key creation certificate chain back to the owner’s paired iPhone to
verify that the key was created on an authentic Apple device. The owner’s paired iPhone
signs the ECC-public key of the other family member’s or friend’s iPhone and sends the
signature back to the family member’s or friend’s iPhone. The signing operation in the
owner device requires user authentication (Touch ID, Face ID or passcode entry) and
a secure user intent described in Uses for Touch ID and Face ID. The authorization is
requested when sending the invitation and is stored in the secure element for consumption
when the friend device sends back the signing request.
Key deletion
Keys can be deleted on the keyholder device from the owner device and in the vehicle.
Deletions on the keyholder iPhone are effective immediately, even if the keyholder uses the
key. Therefore a strong warning is shown before the deletion.
Deletions of keys in the vehicle depends on whether the automaker requires the vehicle to
be online for the deletion or not.
In both cases, the deletion on keyholder device or vehicle is reported to a key inventory
server (KIS) on the automaker side, which registers issued keys for a vehicle for insurance
purposes.
The owner can request a deletion from the back of the owner pass. The request is first sent
to the automaker for key removal in the vehicle. The conditions for removing the key from
the vehicle are defined by the automaker. Only when the key is removed in the vehicle will
the automaker server send a remote termination request to the keyholder device.
When a key is terminated in a device, the applet that manages the digital car keys creates
a cryptographically signed termination attestation, which is used as proof of deletion by
the automaker and used to remove the key from the KTS.
Standard transactions
A secure channel between the reader and an iPhone is initiated by generating ephemeral
key pairs on the reader and the iPhone side. Using a key agreement method, a shared
secret can be derived on both sides and used for generation of a shared symmetric key
using Diffie-Hellman, a key derivation function, and signatures from the long-term key
established during pairing.
The ephemeral public key generated on the vehicle side is signed with the reader’s long-
term private key, which results in an authentication of the reader by the iPhone. From the
iPhone perspective, this protocol is designed to prevent privacy-sensitive data from being
revealed to an adversary intercepting the communication.
Finally, the iPhone uses the established secure channel to encrypt its public key identifier
along with the signature computed on a reader’s data-derived challenge and some
additional app-specific data. This verification of the iPhone signature by the reader allows
the reader to authenticate the device.

Fast transactions
The iPhone generates a cryptogram based on a secret previously shared during a standard
transaction. This cryptogram allows the vehicle to quickly authenticate the device in
performance sensitive scenarios. Optionally, a secure channel between the vehicle and the
device is established by deriving session keys from a secret previously shared during a
standard transaction and a new ephemeral key pair. The ability of the vehicle to establish
the secure channel authenticates the vehicle to the iPhone.
Privacy
The key tracking server of the automaker doesn’t store the device ID, SEID, or Apple ID. It
stores only a mutable identifier—the instance CA identifier. This identifier isn’t bound to
any private data in the device or by the server, and it’s deleted when the user wipes their
device completely (using Erase All Contents and Settings).

Network security
Network security overview

In addition to the built-in safeguards Apple uses to protect data stored on Apple devices,
there are many measures organizations can take to keep information secure as it travels to
and from a device. All of these safeguards and measures fall under network security.
Users must be able to access corporate networks from anywhere in the world, so it’s
important to ensure that they are authorized and that their data is protected during
transmission. To accomplish these security objectives, iOS, iPadOS, and macOS
integrate proven technologies and the latest standards for both Wi-Fi and cellular data
network connections. That’s why our operating systems use—and provide developer
access to—standard networking protocols for authenticated, authorized, and encrypted
communications.
TLS security
iOS, iPadOS, and macOS support Transport Layer Security (TLS 1.0, TLS 1.1, TLS 1.2, TLS
1.3) and Datagram Transport Layer Security (DTLS). The TLS protocol supports both
AES128 and AES256, and prefers cipher suites with forward secrecy. Internet apps such
as Safari, Calendar, and Mail automatically use this protocol to enable an encrypted
communication channel between the device and network services. High-level APIs (such
as CFNetwork) make it easy for developers to adopt TLS in their apps, while low-level APIs
(such as Network.framework) provide fine-grained control. CFNetwork disallows SSL 3, and
apps that use WebKit (such as Safari) are prohibited from making an SSL 3 connection.
In iOS 11 or later and macOS 10.13 or later, SHA-1 certificates are no longer allowed
for TLS connections unless trusted by the user. Certificates with RSA keys shorter than
2048 bits are also disallowed. The RC4 symmetric cipher suite is deprecated in iOS 10
and macOS 10.12. By default, TLS clients or servers implemented with SecureTransport
APIs don’t have RC4 cipher suites enabled and are unable to connect when RC4 is
the only cipher suite available. To be more secure, services or apps that require RC4
should be upgraded to use secure cipher suites. In iOS 12.1, certificates issued after
October 15, 2018, from a system-trusted root certificate must be logged in a trusted
Certificate Transparency log to be allowed for TLS connections. In iOS 12.2, TLS 1.3 is
enabled by default for Network.framework and NSURLSession APIs. TLS clients using the
SecureTransport APIs can’t use TLS 1.3.

App Transport Security
App Transport Security provides default connection requirements so that apps adhere
to best practices for secure connections when using NSURLConnection, CFURL, or
NSURLSession APIs. By default, App Transport Security limits cipher selection to include
only suites that provide forward secrecy, specifically:
• ECDHE_ECDSA_AES and ECDHE_RSA_AES in Galois/Counter Mode (GCM)
• Cipher Block Chaining (CBC) mode
Apps are able to disable the forward secrecy requirement per domain, in which case RSA_
AES is added to the set of available ciphers.
Servers must support TLS 1.2 and forward secrecy, and certificates must be valid and
signed using SHA256 or stronger with a minimum 2048-bit RSA key or 256-bit elliptic
curve key.
Network connections that don’t meet these requirements will fail unless the app
overrides App Transport Security. Invalid certificates always result in a hard failure and no
connection. App Transport Security is automatically applied to apps that are compiled for
iOS 9 or later and macOS 10.11 or later.
Certificate validity checking

Evaluating the trusted status of a TLS certificate is performed in accordance with
established industry standards, as set out in RFC 5280, and incorporates emerging
standards such as RFC 6962 (Certificate Transparency). In iOS 11 or later and macOS 10.13
or later, Apple devices are periodically updated with a current list of revoked and
constrained certificates. The list is aggregated from certificate revocation lists (CRLs),
which are published by each of the built-in root certificate authorities trusted by Apple,
as well as by their subordinate CA issuers. The list may also include other constraints
at Apple’s discretion. This information is consulted whenever a network API function is
used to make a secure connection. If there are too many revoked certificates from a CA
to list individually, a trust evaluation may instead require that an online certificate status
response (OCSP) is needed, and if the response isn’t available, the trust evaluation will fail.

IPv6 security
All Apple operating systems support IPv6, implementing several mechanisms to protect
the privacy of users and the stability of the networking stack. When Stateless Address
Autoconfiguration (SLAAC) is used, the IPv6 addresses of all interfaces are generated in a
way that prevents tracking devices across networks and at the same time allows for a good
user experience by ensuring address stability when no network changes take place. The
address generation algorithm is based on cryptographically generated addresses as of RFC
3972, enhanced by an interface-specific modifier to warrant that even different interfaces
on the same network eventually have different addresses. Furthermore, temporary
addresses are created with a preferred lifetime of 24 hours, and these are used by default
for any new connections. Aligned with the Private Wi-Fi address feature introduced in
iOS 14, iPadOS 14, and watchOS 7,a unique link-local address is generated for every Wi-Fi
network that a device joins. The network’s SSID is incorporated as an additional element
for the address generation, similar to the Network_ID parameter as of RFC 7217. This
approach is used in iOS 14, iPadOS 14, and watchOS 7.
To protect against attacks based on IPv6 extension headers and fragmentation, Apple
devices implement protection measures specified in RFC 6980, RFC 7112, and RFC 8021.
Among other measures, these inhibit attacks where the upper-layer header can be found
only in the second fragment (as shown below), which in turn could cause ambiguities for
security controls like stateless packet filters.
An IPv6 datagram.
In addition, to ensure the reliability of the iPv6 stack of Apple operating systems, Apple
devices enforce various limits on IPv6-related data structures, such as the number of
prefixes per interface.

Virtual private network (VPN) security
Secure network services like virtual private networking typically require minimal setup and
configuration to work with iOS, iPadOS, and macOS devices.
Protocols supported
These devices work with VPN servers that support the following protocols and
authentication methods:
• IKEv2/IPsec with authentication by shared secret, RSA Certificates, Elliptic Curve Digital
Signature Algorithm (ECDSA) Certificates, EAP-MSCHAPv2, or EAP-TLS
• SSL-VPN using the appropriate client app from the App Store
• L2TP/IPsec with user authentication by MS-CHAPV2 password and machine

authentication by shared secret (iOS, iPadOS, and macOS) and RSA SecurID or
CRYPTOCard (macOS only)
• Cisco IPsec with user authentication by password, RSA SecurID or CRYPTOCard, and
machine authentication by shared secret and certificates (macOS only)
VPN deployments supported

iOS, iPadOS, and macOS support the following:
• VPN On Demand: For networks that use certificate-based authentication. IT policies

specify which domains require a VPN connection by using a VPN configuration profile.
• Per App VPN: For facilitating VPN connections on a much more granular basis. Mobile
device management (MDM) solutions can specify a connection for each managed app
and specific domains in Safari. This helps ensure that secure data always goes to and
from the corporate network—and that a user’s personal data doesn’t.
• Always On VPN: Can be configured for devices managed through an MDM solution
and supervised using Apple Configurator 2, Apple School Manager, or Apple Business
Manager. Always On VPN eliminates the need for users to turn on VPN to enable
protection when connecting to cellular and Wi-Fi networks. Always On VPN gives
an organization full control over device traffic by tunneling all IP traffic back to the
organization. The default exchange of parameters and keys for the subsequent
encryption, IKEv2 secures traffic transmission with data encryption. The organization
can monitor and filter traffic to and from its devices, secure data within its network, and
restrict device access to the internet.

Wi-Fi security
Protocol security
Secure access to wireless networks

All Apple platforms support industry-standard Wi-Fi authentication and encryption
protocols, to provide authenticated access and confidentiality when connecting to the
following secure wireless networks:
• WPA2 Personal
• WPA2 Enterprise
• WPA2/WPA3 Transitional
• WPA3 Personal
• WPA3 Enterprise
• WPA3 Enterprise 192-bit security
WPA2 and WPA3 authenticate each connection and provide 128-bit AES encryption to
ensure confidentiality for data sent over the air. This grants users the highest level of
assurance that their data remains protected when sending and receiving communications
over a Wi-Fi network connection.
WPA3 support
WPA3 is supported on the following Apple devices:
• iPhone 7 or later
• iPad 5th generation or later
• Apple TV 4K or later
• Apple Watch series 3 or later
• Mac computers (late 2013 on, with 802.11ac or later)
Newer devices support authentication with WPA3 Enterprise 192-bit security, including
support for 256-bit AES encryption when connecting to compatible wireless access
points (APs). This provides even stronger confidentiality protections for traffic sent
over the air. WPA3 Enterprise 192-bit security is supported on iPhone 11, iPhone 11 Pro,
iPhone 11 Pro Max, and later iOS and iPadOS devices.

PMF support
In addition to protecting data sent over the air, Apple platforms extend WPA2 and WPA3
level protections to unicast and multicast management frames through the Protected
Management Frame (PMF) service defined in 802.11w. PMF support is available on the
following Apple devices:
• iPad Air 2 or later
• Apple TV HD or later
• Mac computers (late 2013 on, with 802.11ac or later)
With support for 802.1X, Apple devices can be integrated into a broad range of RADIUS
authentication environments. 802.1X wireless authentication methods supported include
EAP-TLS, EAP-TTLS, EAP-FAST, EAP-SIM, PEAPv0, and PEAPv1.
Platform protections
Apple operating systems protect the device from vulnerabilities in network processor
firmware. This means that network controllers with Wi-Fi have limited access to Application
Processor memory.
• When USB or SDIO (Secure Digital Input Output) is used to interface with the
network processor, the network processor can’t initiate direct memory access (DMA)
transactions to the Application Processor.
• When PCIe is used, each network processor is on its own isolated PCIe bus. An Input/
Output Memory Management Unit (IOMMU) on each PCIe bus further limits the network
processor’s DMA access to only memory and resources containing its network packets
and control structures.
Deprecated protocols
Apple products support the following deprecated Wi-Fi authentication and encryption
protocols:
• WEP Open, with both 40-bit and 104-bit keys
• WEP Shared, with both 40-bit and 104-bit keys
• Dynamic WEP
• Temporal Key Integrity Protocol (TKIP)
• WPA
• WPA/WPA2 Transitional
These protocols are no longer considered secure, and their use is strongly discouraged
for compatibility, reliability, performance, and security reasons. They are supported for
backward compatibility purposes only and may be removed in future software versions.
It’s recommended that all Wi-Fi implementations be migrated to WPA3 Personal or WPA3
Enterprise, to provide the most robust, secure, and compatible Wi-Fi connections possible.

Wi-Fi privacy
MAC address randomization

Apple platforms use a randomized media access control address (MAC address) when
performing Wi-Fi scans when not associated with a Wi-Fi network. These scans can be
performed to find and connect to a known Wi-Fi network or to assist Location Services for
apps that use geofences, such as location-based reminders or fixing a location in Apple
Maps. Note that Wi-Fi scans that happen while trying to connect to a preferred Wi-Fi
network aren’t randomized. Wi-Fi MAC address randomization support is available on
iPhone 5 or later.
Apple platforms also use a randomized MAC address when conducting enhanced Preferred
Network Offload (ePNO) scans when a device isn’t associated with a Wi-Fi network or its
processor is asleep. ePNO scans are run when a device uses Location Services for apps
that use geofences, such as location-based reminders that determine whether the device
is near a specific location.
Because a device’s MAC address changes when disconnected from a Wi-Fi network,
it can’t be used to persistently track a device by passive observers of Wi-Fi traffic,
even when the device is connected to a cellular network. Apple has informed Wi-Fi
manufacturers that iOS and iPadOS Wi-Fi scans use a randomized MAC address and that
neither Apple nor manufacturers can predict these randomized MAC addresses.
iOS 14, iPadOS 14, and watchOS 7 introduce a new Wi-Fi privacy feature: When an iPhone,
iPad, iPod touch, or Apple Watch connects to a Wi-Fi network, it identifies itself with a a
unique (random) MAC address per network. This feature can be disabled either by the user
or using a new option in the Wi-Fi payload. Under certain circumstances, the device will
fall back to the actual MAC address.
For more information, see the Apple Support article Use private Wi-Fi addresses in iOS 14,
iPadOS 14, and watchOS 7.

Wi-Fi frame sequence number randomization
Wi-Fi frames include a sequence number, which is used by the low-level 802.11 protocol
to enable efficient and reliable Wi-Fi communications. Because these sequence numbers
increment on each transmitted frame, they could be used to correlate information
transmitted during Wi-Fi scans with other frames transmitted by the same device.
To guard against this, Apple devices randomize the sequence numbers whenever a MAC
address is changed to a new randomized address. This includes randomizing the sequence
numbers for each new scan request that’s initiated while the device is unassociated. This
randomization is supported on the following devices:
• iPad 5th generation or later
• Apple TV 4K or later
• iMac Pro (Retina 5K, 27-inch, 2017) or later
• MacBook Pro (13-inch, 2018) or later
• MacBook Pro (15-inch, 2018) or later
• MacBook Air (Retina, 13-inch, 2018) or later
• Mac mini (2018) or later
• iMac (Retina 4K, 21.5-inch, 2019) or later
• iMac (Retina 5K, 27-inch, 2019) or later
• Mac Pro (2019) or later
Wi-Fi connections
Apple generates randomized MAC addresses for the Peer-to-Peer Wi-Fi connections
that are used for AirDrop and AirPlay. Randomized addresses are also used for Personal
Hotspot in iOS and iPadOS (with a SIM card) and Internet Sharing in macOS.
New random addresses are generated whenever these network interfaces are started, and
unique addresses are independently generated for each interface as needed.
Hidden networks
Wi-Fi networks are identified by their network name, known as a service set identifier
(SSID). Some Wi-Fi networks are configured to hide their SSID, which results in the
wireless access point not broadcasting the network’s name. These are known as hidden
networks. iPhone 6s or later automatically detects when a network is hidden. If a network
is hidden, the iOS or iPadOS device sends a probe with the SSID included in the request—
not otherwise. This prevents the device from broadcasting the name of previously hidden
networks a user was connected to, thereby further ensuring privacy.

Bluetooth security
There are two types of Bluetooth in Apple devices, Bluetooth Classic and Bluetooth Low
Energy (BLE). The Bluetooth security model for both versions includes the following
distinct security features:
• Pairing: The process for creating one or more shared secret keys
• Bonding: The act of storing the keys created during pairing for use in subsequent
connections to form a trusted device pair
• Authentication: Verifying that the two devices have the same keys
• Encryption: Message confidentiality
• Message integrity: Protection against message forgeries
• Secure Simple Pairing: Protection against passive eavesdropping and protection against
man-in-the-middle attacks
Bluetooth version 4.1 added the Secure Connections feature to Bluetooth Classic (BR/EDR)
physical transport.
The security features for each type of Bluetooth are listed below.
Support Bluetooth Classic Bluetooth Low Energy
Pairing P-256 elliptic curve FIPS-approved algorithms

(AES-CMAC and P-256 elliptic curve)
Bonding Pairing information stored in Pairing information stored in a secure

a secure location in iO S, iPad location in iO S, iPadO S, macOS,
OS, macO S, tvO S, and watchO S tvO S, and watchO S devices
devices
Authentication FIPS-approved algorithms FIPS-approved algorithms

(HMAC-SHA256 and AES-CTR)
Encryption AES-CCM cryptography, performed AES-CCM cryptography, performed in

in the Controller the Controller
Message integrity AES-CCM, used for message AES-CCM, used for message integrity
integrity
Secure Simple Pairing: Protection Elliptic Curve Diffie-Hellman Elliptic Curve Diffie-Hellman
against passive eavesdropping Exchange (ECDHE) Exchange (ECDHE)
Secure Simple Pairing: Protection Two user-assisted numeric Two user-assisted numeric methods:
against man-in-the-middle (MITM) methods: numerical comparison or numerical comparison or passkey
attacks passkey entry entry
Pairings require a user response,

including all non-MITM pairing modes
Bluetooth 4.1 or later iMac Late 2015 or later iO S 9 or later
MacB ook Pro Early 2015 or later iPadO S 13.1 or later
macO S 10.12 or later
tvO S 9 or later
watchO S 2.0 or later

Support Bluetooth Classic Bluetooth Low Energy
Bluetooth 4.2 or later iPhone 6 or later iO S 9 or later
iPadO S 13.1 or later
macO S 10.12 or later
tvO S 9 or later
watchO S 2.0 or later
Bluetooth Low Energy privacy

To help secure user privacy, BLE includes the following two features: address
randomization and cross-transport key derivation.
Address randomization is a feature that reduces the ability to track a BLE device over a
period of time by changing the Bluetooth device address on a frequent basis. For a device
using the privacy feature to reconnect to known devices, the device address, referred
to as the private address, must be resolvable by the other device. The private address
is generated using the device’s identity resolving key exchanged during the pairing
procedure.
iOS 13 or later and iPadOS 13.1 or later have the ability to derive link keys across
transports, a feature known as cross-transport key derivation. For example, a link key
generated with BLE can be used to derive a Bluetooth Classic link key. In addition, Apple
added Bluetooth Classic to BLE support for devices that support the Secured Connections
feature that was introduced in the Bluetooth Core Specification 4.1 (see the Bluetooth Core
Specification 5.1).
Ultra Wideband security in iOS

The new Apple‑designed U1 chip uses Ultra Wideband technology for spatial awareness—
allowing iPhone 11, iPhone 11 Pro, and iPhone 11 Pro Max or later iPhone models to
precisely locate other U1‑equipped Apple devices. Ultra Wideband technology uses the
same technology to randomize data found in other supported Apple devices:
• MAC address randomization

• Wi-Fi frame sequence number randomization

Single sign-on security
Single sign-on
iOS and iPadOS support authentication to enterprise networks through Single sign-on
(SSO). SSO works with Kerberos-based networks to authenticate users to services they
are authorized to access. SSO can be used for a range of network activities, from secure
Safari sessions to third-party apps. Certificate-based authentication such as PKINIT is also
supported.
macOS supports authentication to enterprise networks using Kerberos. Apps can use
Kerberos to authenticate users to services they’re authorized to access. Kerberos can also
be used for a range of network activities, from secure Safari sessions and network file
system authentication to third-party apps. Certificate-based authentication is supported,
although app adoption of a developer API is required.
iOS, iPadOS, and macOS SSO use SPNEGO tokens and the HTTP Negotiate protocol to
work with Kerberos-based authentication gateways and Windows Integrated Authentication
systems that support Kerberos tickets. SSO support is based on the open source Heimdal
project.
The following encryption types are supported in iOS, iPadOS, and macOS:
• AES-128-CTS-HMAC-SHA1-96
• AES-256-CTS-HMAC-SHA1-96
• DES3-CBC-SHA1
• ARCFOUR-HMAC-MD5
Safari supports SSO, and third-party apps that use standard iOS and iPadOS networking
APIs can also be configured to use it. To configure SSO, iOS and iPadOS support a
configuration profile payload that allows mobile device management (MDM) solutions to
push down the necessary settings. This includes setting the user principal name (that
is, the Active Directory user account) and Kerberos realm settings, as well as configuring
which apps and Safari web URLs should be allowed to use SSO.
To configure Kerberos in macOS, acquire tickets with Ticket Viewer, log in to a Windows
Active Directory domain, or use the kinit command-line tool.
Extensible Single sign-on

App developers can provide their own single sign-on implementations using SSO
extensions. SSO extensions are invoked when a native or web app needs to use some
identity provider for user authentication. Developers can provide two types of extensions:
those that redirect to HTTPS and those that use a challenge/response mechanism such as
Kerberos. This allows OpenID, OAuth, SAML2 and Kerberos authentication schemes to be
supported by Extensible Single sign-on.

To use a Single sign-on extension, an app can either use the AuthenticationServices API or
can rely on the URL interception mechanism offered by the operating system. WebKit and
CFNetwork provide an interception layer that enables a seamless support of Single sign-on
for any native or WebKit app. For a Single sign-on extension to be invoked, a configuration
provided by an administrator has to be installed though a mobile device management
(MDM) profile. In addition, redirect type extensions must use the Associated Domains
payload to prove that the identity server they support is aware of their existence.
The only extension provided with the operating system is the Kerberos SSO extension.
AirDrop security
Apple devices that support AirDrop use Bluetooth Low Energy (BLE) and Apple-created
peer-to-peer Wi-Fi technology to send files and information to nearby devices, including
AirDrop-capable iOS devices running iOS 7 or later and Mac computers running OS X 10.11
or later. The Wi-Fi radio is used to communicate directly between devices without using
any internet connection or wireless access point (AP). In macOS, this connection is
encrypted with TLS.
AirDrop is set to share with Contacts Only by default. Users can also choose to use AirDrop
to share with everyone, or turn off the feature entirely. Organizations can restrict the use of
AirDrop for devices or apps being managed by using a mobile device management (MDM)
solution.
AirDrop operation
AirDrop uses iCloud services to help users authenticate. When a user signs into iCloud,
a 2048-bit RSA identity is stored on the device, and when the user enables AirDrop, an
AirDrop short identity hash is created based on the email addresses and phone numbers
associated with the user’s Apple ID.
When a user chooses AirDrop as the method for sharing an item, the sending device emits
an AirDrop signal over BLE that includes the user’s AirDrop short identity hash. Other
Apple devices that are awake, in close proximity, and have AirDrop turned on, detect the
signal and respond using peer-to-peer Wi-Fi, so that the sending device can discover the
identity of any responding devices.
In Contacts Only mode, the received AirDrop short identity hash is compared with hashes
of people in the receiving device’s Contacts app. If a match is found, the receiving device
responds over peer-to-peer Wi-Fi with its identity information. If there is no match, the
device doesn’t respond.
In Everyone mode, the same overall process is used. However, the receiving device
responds even if there is no match in the device’s Contacts app.
The sending device then initiates an AirDrop connection using peer-to-peer Wi-Fi, using
this connection to send a long identity hash to the receiving device. If the long identity
hash matches the hash of a known person in the receiver’s Contacts, then the receiver
responds with its long identity hashes.

If the hashes are verified, the recipient’s first name and photo (if present in Contacts) are
displayed in the sender’s AirDrop share sheet. In iOS and iPadOS, they are shown in the
“People” or “Devices” section. Devices that aren’t verified or authenticated are displayed in
the sender’s AirDrop share sheet with a silhouette icon and the device’s name, as defined
in Settings > General > About > Name. In iOS and iPadOS, they are placed in the “Other
People” section of the AirDrop share sheet.
The sending user may then select whom they want to share with. Upon user selection, the
sending device initiates an encrypted (TLS) connection with the receiving device, which
exchanges their iCloud identity certificates. The identity in the certificates is verified
against each user’s Contacts app.
If the certificates are verified, the receiving user is asked to accept the incoming transfer
from the identified user or device. If multiple recipients have been selected, this process is
repeated for each destination.
Wi-Fi password sharing security on iPhone and

iPad
iOS and iPadOS devices that support Wi-Fi password sharing use a mechanism similar to
AirDrop to send a Wi-Fi password from one device to another.
When a user selects a Wi-Fi network (requestor) and is prompted for the Wi-Fi password,
the Apple device starts a Bluetooth Low Energy (BLE) advertisement indicating that it
wants the Wi-Fi password. Other Apple devices that are awake, in close proximity, and
have the password for the selected Wi-Fi network connect using BLE to the requesting
device.
The device that has the Wi-Fi password (grantor) requires the Contact information of
the requestor, and the requestor must prove their identity using a similar mechanism to
AirDrop. After identity is proven, the grantor sends the requestor the passcode which can
be used to join the network.
Organizations can restrict the use of Wi-Fi password sharing for devices or apps being
managed through a mobile device management (MDM) solution.

Firewall security in macOS
macOS includes a built-in firewall to protect the Mac from network access and denial-of-
service attacks. It can be configured in the Security & Privacy pane of System Preferences
and supports the following configurations:
• Block all incoming connections, regardless of app.
• Automatically allow built-in software to receive incoming connections.
• Automatically allow downloaded and signed software to receive incoming connections.
• Add or deny access based on user-specified apps.
• Prevent the Mac from responding to ICM (Internet Control Message Protocol) probing
and portscan requests.

Developer kit security
Developer kit security overview

Apple provides a number of “kit” frameworks to enable third-party developers to extend
Apple services. These frameworks are built with user security and privacy at their core:
• HomeKit
• CloudKit
• SiriKit
• DriverKit
• ReplayKit
• ARKit
HomeKit
HomeKit communication security
Overview
HomeKit provides a home automation infrastructure that uses iCloud and iOS, iPadOS, and
macOS security to protect and sync private data without exposing it to Apple.
HomeKit identity and security are based on Ed25519 public-private key pairs. An Ed25519
key pair is generated on the iOS, iPadOS, and macOS device for each user for HomeKit,
which becomes their HomeKit identity. It’s used to authenticate communication between
iOS, iPadOS, and macOS devices, and between iOS, iPadOS, and macOS devices and
accessories.
The keys—stored in keychain and are included only in encrypted Keychain backups—are
kept up to date between devices using iCloud Keychain, where available. HomePod and
Apple TV receive keys using tap-to-setup or the setup mode described below. Keys are
shared from an iPhone to a paired Apple Watch using Apple Identity Service (IDS).
Communication between HomeKit accessories

HomeKit accessories generate their own Ed25519 key pair for use in communicating with
iOS, iPadOS, and macOS devices. If the accessory is restored to factory settings, a new
key pair is generated.

To establish a relationship between an iOS, iPadOS, and macOS device and a HomeKit
accessory, keys are exchanged using Secure Remote Password (3072-bit) protocol
utilizing an eight-digit code provided by the accessory’s manufacturer, entered on the
iOS, iPadOS device by the user, and then encrypted using ChaCha20-Poly1305 AEAD with
HKDF-SHA512 derived keys. The accessory’s MFi certification is also verified during setup.
Accessories without an MFi chip can build in support for software authentication in iOS 11.3
or later.
When the iOS, iPadOS, and macOS device and the HomeKit accessory communicate during
use, each authenticates the other using the keys exchanged in the above process. Each
session is established using the Station-to-Station protocol and is encrypted with HKDF-
SHA512 derived keys based on per-session Curve25519 keys. This applies to both IP-
based and Bluetooth Low Energy (BLE) accessories.
For BLE devices that support broadcast notifications, the accessory is provisioned with
a broadcast encryption key by a paired iOS, iPadOS, and macOS device over a secure
session. This key is used to encrypt the data about state changes on the accessory, which
are notified using the BLE advertisements. The broadcast encryption key is an HKDF-
SHA512 derived key, and the data is encrypted using ChaCha20-Poly1305 AEAD algorithm.
The broadcast encryption key is periodically changed by the iOS, iPadOS, and macOS
device and updated to other devices using iCloud as described in HomeKit data security.
HomeKit and Siri

Siri can be used to query and control accessories, and to activate scenes. Minimal
information about the configuration of the home is provided anonymously to Siri, to provide
names of rooms, accessories, and scenes that are necessary for command recognition.
Audio sent to Siri may denote specific accessories or commands, but such Siri data isn’t
associated with other Apple features such as HomeKit.
HomeKit data security
HomeKit data updates between devices and users

HomeKit data can be updated between a user’s iOS, iPadOS, and macOS devices using
iCloud and iCloud keychain. During this process, the HomeKit data is encrypted using keys
derived from the user’s HomeKit identity and a random nonce and is handled as an opaque
binary large object, or blob. The most recent blob is stored in iCloud, but it isn’t used for
any other purpose. Because it’s encrypted using keys that are available only on the user’s
iOS, iPadOS, and macOS devices, its contents are inaccessible during transmission and
iCloud storage.
HomeKit data is also synced between multiple users of the same home. This process uses
authentication and encryption that is the same as that used between an iOS, iPadOS, and
macOS device and a HomeKit accessory. The authentication is based on Ed25519 public
keys that are exchanged between the devices when a user is added to a home. After a new
user is added to a home, all further communication is authenticated and encrypted using
Station-to-Station protocol and per-session keys.

The user who initially created the home in HomeKit or another user with editing
permissions can add new users. The owner’s device configures the accessories with the
public key of the new user so that the accessory can authenticate and accept commands
from the new user. When a user with editing permissions adds a new user, the process is
delegated to a home hub to complete the operation.
The process to provision Apple TV for use with HomeKit is performed automatically when
the user signs in to iCloud. The iCloud account needs to have two-factor authentication
enabled. Apple TV and the owner’s device exchange temporary Ed25519 public keys
over iCloud. When the owner’s device and Apple TV are on the same local network, the
temporary keys are used to secure a connection over the local network using Station-to-
Station protocol and per-session keys. This process uses authentication and encryption
that is the same as that used between an iOS, iPadOS, and macOS device and a HomeKit
accessory. Over this secure local connection, the owner’s device transfers the user’s
Ed25519 public-private key pairs to Apple TV. These keys are then used to secure the
communication between Apple TV and the HomeKit accessories and also between
Apple TV and other iOS, iPadOS, and macOS devices that are part of the HomeKit home.
If a user doesn’t have multiple devices and doesn’t grant additional users access to their
home, no HomeKit data is transmitted to iCloud.
Home data and apps

Access to home data by apps is controlled by the user’s Privacy settings. Users are asked
to grant access when apps request home data, similar to Contacts, Photos, and other iOS,
iPadOS, and macOS data sources. If the user approves, apps have access to the names of
rooms, names of accessories, which room each accessory is in, and other information as
detailed in the HomeKit developer documentation at https://developer.apple.com/homekit/.
Local data storage

HomeKit stores data about the homes, accessories, scenes, and users on a user’s iOS,
iPadOS, and macOS devices. This stored data is encrypted using keys derived from the
user’s HomeKit identity keys, plus a random nonce. Additionally, HomeKit data is stored
using the Data Protection class Protected Until First User Authentication. HomeKit data
is backed up only in encrypted backups, so, for example, unencrypted backups to iTunes
(in macOS 10.14 or earlier) or the Finder (macOS 10.15 or later) through USB don’t contain
HomeKit data.

Securing routers with HomeKit
Routers that support HomeKit let users improve the security of their home network by
managing the Wi-Fi access that HomeKit accessories have to their local network and to the
internet. The routers also support Private PSK (PPSK) authentication, so accessories can
be added to the Wi-Fi network using a key that’s specific to the accessory and that can be
revoked when needed. PPSK authentication improves security by not exposing the main
Wi-Fi password to accessories, as well as by allowing the router to securely identify an
accessory even if it were to change its MAC address.
Using the Home app, a user can configure access restrictions for groups of accessories as
follows:
• No restriction: Allow unrestricted access to the internet and the local network.
• Automatic: This is the default setting. Allow access to the internet and the local network
based on a list of internet sites and local ports provided to Apple by the accessory
manufacturer. This list includes all sites and ports needed by the accessory in order to
function properly. (No Restriction is in place until such a list is available.)
• Restrict to Home: No access to the internet or the local network except for the
connections required by HomeKit to discover and control the accessory from the local
network (including from the home hub to support remote control).
A PPSK is a strong, accessory-specific WPA2 Personal pass-phrase that is automatically

generated by HomeKit and revoked if and when the accessory is later removed from the
Home. A PPSK is used when an accessory is added to the Wi-Fi network by HomeKit
in a Home that has been configured with a HomeKit router; this addition is reflected
as Wi-Fi Credential: HomeKit-managed on the settings screen for the accessory in the
Home app. Accessories that were added to the Wi-Fi network before adding the router
are reconfigured to use a PPSK if the accessory supports this; otherwise, they retain their
existing credentials.
As an additional security measure, users must configure the HomeKit router using the
router manufacturer’s app, so that the app can validate that users have access to the
router and can add it to the Home app.
HomeKit camera security

Cameras that have an Internet Protocol address (IP address) in HomeKit send video and
audio streams directly to the iOS, iPadOS, tvOS, and macOS device on the local network
accessing the stream. The streams are encrypted using randomly generated keys on the
device and an Internet Protocol camera (or IP camera), and they’re exchanged over the
secure HomeKit session to the camera. When a device isn’t on the local network, the
encrypted streams are relayed through the home hub to the device. The home hub doesn’t
decrypt the streams; it functions only as a relay between the device and the IP camera.
When an app displays the HomeKit IP camera video view to the user, HomeKit renders the
video frames securely from a separate system process. As a result, the app is unable to
access or store the video stream. In addition, apps aren’t permitted to capture screenshots
from this stream.

HomeKit secure video
HomeKit provides an end-to-end secure and private mechanism to record, analyze, and
view clips from HomeKit IP cameras without exposing that video content to Apple or any
third party. When motion is detected by the IP camera, video clips are sent directly to an
Apple device acting as a home hub, using a dedicated local network connection between
that home hub and the IP camera. The local network connection is encrypted with a
per-session HKDF-SHA512 derived key-pair that is negotiated over the HomeKit session
between home hub and IP camera. HomeKit decrypts the audio and video streams on the
home hub and analyzes the video frames locally for any significant event. If a significant
event is detected, HomeKit encrypts the video clip using AES-256-GCM with a randomly
generated AES256 key. HomeKit also generates poster frames for each clip and these
poster frames are encrypted using the same AES256 key. The encrypted poster frame and
audio and video data are uploaded to iCloud servers. The related metadata for each clip
including the encryption key are uploaded to CloudKit using iCloud end-to-end encryption.
For face classification, HomeKit stores all data used to classify a particular person’s face in
CloudKit using iCloud end-to-end encryption. The data stored includes information about
each person, such as name, as well as images representing that person’s face. These face
images can be sourced from a user’s Photos if they opt in, or they can be collected from
previously analyzed IP camera video. A HomeKit Secure Video analysis session uses this
classification data to identify faces in the secure video stream it receives directly from
the IP camera and includes that identification information in the clip metadata mentioned
previously.
When the Home app is used to view the clips for a camera, the data is downloaded from
iCloud and the keys to decrypt the streams are unwrapped locally using iCloud end-to-
end decryption. The encrypted video content is streamed from the servers and decrypted
locally on the iOS device before displaying it in the viewer. Each video clip session maybe
broken down into sub-sections with each sub-section encrypting the content stream with
its own unique key.
HomeKit security with Apple TV
Using third-party remote accessories with Apple TV

Some third-party remote accessories provide Human Interface Design (HID) events and
Siri audio to an associated Apple TV added using the Home app. The remote sends the
HID events over the secure session to the Apple TV. A Siri-capable TV remote sends audio
data to Apple TV when the user explicitly activates the microphone on the remote using a
dedicated Siri button. The remote sends the audio frames directly to the Apple TV using
a dedicated local network connection. A per-session HKDF-SHA512 derived key-pair that
is negotiated over the HomeKit session between Apple TV and the TV remote is used to
encrypt the local network connection. HomeKit decrypts the audio frames on Apple TV and
forwards them to the Siri app, where they are treated with the same privacy protections as
all Siri audio input.

Apple TV profiles for HomeKit homes
When a user of a HomeKit home adds their profile to the owner of the home’s Apple TV,
it gives that user access to their TV shows, music, and podcasts. Settings for each user
regarding their profile use on the Apple TV are shared to the owner’s iCloud account using
iCloud end-to-end encryption. The data is owned by each user and is shared as read-only
to the owner. Each user of the home can change these values in the Home app and the
Apple TV of the owner uses these settings.
When a setting is turned on, the iTunes account of the user is made available on the
Apple TV. When a setting is turned off, all account and data pertaining to that user is
deleted on the Apple TV. The initial CloudKit share is initiated by the user’s device and the
token to establish the secure CloudKit share is sent over the same secure channel that is
used to sync data between users of the home.
HomeKit accessories and iCloud

Note: Whenever possible, access services directly through a home hub instead of through
iCloud. For example, use a hub such as HomePod, Apple TV, or iPad.
iCloud remote access is still supported for legacy HomeKit devices. Apple carefully
designed those devices so that users can control them and send notifications to them
without revealing to Apple what the accessories are or what commands and notifications
are being sent. HomeKit never sends information about the home over iCloud remote
access.
Mutual authentication of an accessory and an Apple device

When a user sends a command using iCloud remote access, the accessory and iOS,
iPadOS, and macOS device are mutually authenticated and data is encrypted using the
same procedure described for local connections. The contents of the communications are
encrypted and not visible to Apple. The addressing through iCloud is based on the iCloud
identifiers registered during the setup process.

Accessory setup process
Accessories that support iCloud remote access are provisioned during the accessory’s
setup process. The provisioning process begins with the user signing in to iCloud. Next,
the iOS, and iPadOS device asks the accessory to sign a challenge using the Apple
Authentication Coprocessor that’s built into all Built for HomeKit accessories. The
accessory also generates prime256v1 elliptic curve keys, and the public key is sent to the
iOS, and iPadOS device along with the signed challenge and the X.509 certificate of the
authentication coprocessor. These are used to request a certificate for the accessory from
the iCloud provisioning server. The certificate is stored by the accessory, but it doesn’t
contain any identifying information about the accessory, other than it has been granted
access to HomeKit iCloud remote access. The iOS, and iPadOS device that is conducting
the provisioning also sends a bag to the accessory, which contains the URLs and other
information needed to connect to the iCloud remote access server. This information isn’t
specific to any user or accessory.
Accessory list of allowed users

Each accessory registers a list of allowed users with the iCloud remote access server.
These users have been granted the ability to control the accessory by the user who added
the accessory to the home. Users are granted an identifier by the iCloud server and can
be mapped to an iCloud account for the purpose of delivering notification messages and
responses from the accessories. Similarly, accessories have iCloud-issued identifiers, but
these identifiers are opaque and don’t reveal any information about the accessory itself.
How accessories connect to iCloud remote access server

When an accessory connects to the HomeKit iCloud remote access server, it presents
its certificate and a pass. The pass is obtained from a different iCloud server, and it
isn’t unique for each accessory. When an accessory requests a pass, it includes its
manufacturer, model, and firmware version in its request. No user-identifying or home-
identifying information is sent in this request. To help protect privacy, connection to the
pass server isn’t authenticated.
Accessories connect to the iCloud remote access server using HTTP/2, secured using
TLS 1.2 with AES128-GCM and SHA256. The accessory keeps its connection to the iCloud
remote access server open so that it can receive incoming messages and send responses
and outgoing notifications to iOS, iPadOS, and macOS devices.

CloudKit security
CloudKit is a framework that lets app developers store key-value data, structured data,
and assets in iCloud. Access to CloudKit is controlled using app entitlements. CloudKit
supports both public and private databases. Public databases are used by all copies of the
app, typically for general assets, and aren’t encrypted. Private databases store the user’s
data.
As with iCloud Drive, CloudKit uses account-based keys to protect the information stored
in the user’s private database and, similar to other iCloud services, files are chunked,
encrypted, and stored using third-party services. CloudKit uses a hierarchy of keys, similar
to Data Protection. The per-file keys are wrapped by CloudKit Record keys. The Record
keys, in turn, are protected by a zone-wide key, which is protected by the user’s CloudKit
Service key. The CloudKit Service key is stored in the user’s iCloud account and is available
only after the user has authenticated with iCloud.
CloudKit end-to-end encryption.
SiriKit security for iOS, iPadOS, and watchOS

Siri uses the app extension system to communicate with third-party apps. On a device,
Siri can access the user’s contact information and the device’s current location. But
before it provides protected data to an app, Siri checks the app’s user-controlled access
permissions. According to those permissions, Siri passes only the relevant fragment of the
original user utterance to the app extension. For example, if an app doesn’t have access
to contact information, Siri won’t resolve a relationship in a user request such as “Pay my
mother 10 dollars using Payment App.” In this case, the app would see only the literal term
“my mother.”
However, if the user has granted the app access to contact information, the app would
receive resolved information about the user’s mother. If a relationship is referenced in the
body portion of a message—for example, “Tell my mother on MessageApp that my brother
is awesome”—Siri doesn’t resolve “my brother” regardless of the app’s permissions.
SiriKit-enabled apps can send app-specific or user-specific vocabulary to Siri, such as

the names of the user’s contacts. This information allows Siri’s speech recognition and
natural language understanding to recognize vocabulary for that app and is associated
with a random identifier. The custom information remains available as long as the identifier
is in use, or until the user disables the app’s Siri integration in Settings, or until the SiriKit-
enabled app is uninstalled.

For an utterance like “Get me a ride to my mom’s home using RideShareApp,” the request
requires location data from the user’s contacts. For that request only, Siri provides the
required information to the app’s extension, regardless of the user permission settings for
location or contact information for the app.
DriverKit security for macOS 10.15

DriverKit is the framework that allows developers to create device drivers that the user
installs on their Mac. Drivers built with DriverKit run in user space, rather than as kernel
extensions, for improved system security and stability. This makes for easier installation
and increases the stability and security of macOS.
The user simply downloads the app (installers aren’t necessary when using system
extensions or DriverKit) and the extension is enabled only when required. These replace
kexts for many use cases, which require administrator privileges to install in /System/
Library or /Library.
IT administrators who use device drivers, cloud storage solutions, networking, and security
apps that require kernel extensions are encouraged to move to newer versions that are
built on system extensions. These newer versions greatly reduce the possibility of kernel
panics on the Mac as well as reduce the attack surface. These new extensions run in the
user space, won’t require special privileges required for installation, and are automatically
removed when the bundling app is moved to the Trash.
The DriverKit framework provides C++ classes for I/O services, device matching, memory
descriptors, and dispatch queues. It also defines I/O-appropriate types for numbers,
collections, strings, and other common types. The user uses these with family-specific
driver frameworks like USBDriverKit and HIDDriverKit. Use the System Extensions
framework to install and upgrade a driver.
ReplayKit security in iOS and iPadOS

ReplayKit is a beta framework that enables developers to add recording and live
broadcasting capabilities to their apps. In addition, it allows users to annotate their
recordings and broadcasts using the device’s front-facing camera and microphone.
Movie recording
There are several layers of security built into recording a movie:
• Permissions dialog: Before recording starts, ReplayKit presents a user consent alert
requesting that the user acknowledge their intent to record the screen, the microphone,
and the front-facing camera. This alert is presented once per app process, and it’s
presented again if the app is left in the background for longer than 8 minutes.
• Screen and audio capture: Screen and audio capture occurs out of the app’s process in
ReplayKit’s daemon replayd. This ensures the recorded content is never accessible to
the app process.
• In-app screen and audio capture: This allows an app to get video and sample buffers,
which is guarded by the permissions dialogue.

• Movie creation and storage: The movie file is written to a directory that’s only
accessible to ReplayKit’s subsystems and is never accessible to any apps. This prevents
recordings being used by third parties without the user’s consent.
• End-user preview and sharing: The user has the ability to preview and share the movie
with a user interface vended by ReplayKit. The user interface is presented out-of-
process through the iOS Extension infrastructure and has access to the generated
movie file.
ReplayKit broadcasting
There are several layers of security built into broadcasting a movie:
• Screen and audio capture: The screen and audio capture mechanism during
broadcasting is identical to movie recording and occurs in replayd.
• Broadcast extensions: For third-party services to participate in ReplayKit broadcasting,

they’re required to create two new extensions that are configured with the com.apple.
broadcast-services endpoint:
• A user interface extension that allows the user to set up their broadcast
• An upload extension that handles uploading video and audio data to the service’s
back-end servers
The architecture ensures that hosting apps have no privileges to the broadcasted video
and audio contents. Only ReplayKit and the third-party broadcast extensions have access.
• Broadcast picker: With the broadcast picker, users initiate system broadcasts directly
from their app using the same system-defined user interface that’s accessible
using Control Center. The user interface is implemented using a private API and is
an extension that lives within the ReplayKit framework. It is out-of-process from the
hosting app.
• Upload extension: The extension that third-party broadcast services implement to

handle video and audio content during broadcasting uses raw unencoded sample
buffers. During this mode of handling, video and audio data is serialized and passed to
the third-party upload extension in real time through a direct XPC connection. Video
data is encoded by extracting the IOSurface object from the video sample buffer,
encoding it securely as an XPC object, sending it over through XPC to the third-party
extension, and decoding it securely back into an IOSurface object.

ARKit security in iOS and iPadOS
ARKit is a framework that lets developers produce augmented reality experiences in their
app or game. Developers can add 2D or 3D elements using the front or rear camera of an
iOS or iPadOS device.
Apple designed cameras with privacy in mind, and third-party apps must obtain the
user’s consent before accessing the camera. In iOS and iPadOS, when a user grants an
app access to their camera, that app can access real-time images from the front and rear
cameras. Apps aren’t allowed to use the camera without transparency that the camera is in
use.
Photos and videos taken with the camera may contain other information, such as where
and when they were taken, the depth of field, and overcapture. If users don’t want photos
and videos taken with the Camera app to include location, they can control this at any time
by going to Settings > Privacy > Location Services > Camera. If users don’t want photos
and video to include location when shared, they can turn location off in the Options menu
in the share sheet.
To better position the user’s AR experience, apps that use ARKit can use world- or face-
tracking information from the other camera. World tracking uses algorithms on the user’s
device to process information from these sensors to determine their position relative to a
physical space. World tracking enables features such as Optical Heading in Maps.

Secure device management
Secure device management overview

iOS, iPadOS, macOS, and tvOS support flexible security policies and configurations that
are easy to enforce and manage. Through them, organizations can protect corporate
information and ensure that employees meet enterprise requirements, even if they are
using devices they’ve provided themselves—for example, as part of a “bring your own
device” (BYOD) program.
Organizations can use resources such as password protection, configuration profiles,

remote wipe, and third-party mobile device management (MDM) solutions to manage fleets
of devices and help keep corporate data secure, even when employees access this data on
their personal devices.
In iOS 13 or later, iPadOS 13.1 or later, and macOS 10.15 or later, Apple devices support a
new user enrollment option specifically designed for BYOD programs. User enrollments
provide more autonomy for users on their own devices, while increasing the security of
enterprise data by storing it on a separate, cryptographically protected APFS (Apple File
System) volume. This provides a better balance of security, privacy, and user experience
for BYOD programs.
Pairing model security for iPhone and iPad

iOS and iPadOS use a pairing model to control access to a device from a host computer.
Pairing establishes a trust relationship between the device and its connected host,
signified by public key exchange. iOS and iPadOS also use this sign of trust to enable
additional functionality with the connected host, such as data syncing. In iOS 9 or later,
services:
• That require pairing can’t be started until after the device has been unlocked by the
user
• Won’t start unless the device has been recently unlocked
• May (such as with photo syncing) require the device to be unlocked to begin

The pairing process requires the user to unlock the device and accept the pairing request
from the host. In iOS 9 or later, the user is also required to enter their passcode, after
which the host and device exchange and save 2048-bit RSA public keys. The host is
then given a 256-bit key that can unlock an escrow keybag stored on the device. The
exchanged keys are used to start an encrypted SSL session, which the device requires
before it sends protected data to the host or starts a service (iTunes or Finder syncing,
file transfers, Xcode development, and so on). To use this encrypted session for all
communication, the device requires connections from a host over Wi-Fi, so it must have
been previously paired over USB. Pairing also enables several diagnostic capabilities. In
iOS 9, if a pairing record hasn’t been used for more than 6 months, it expires. In iOS 11 or
later, this time frame is shortened to 30 days.
Certain diagnostic services, including com.apple.mobile.pcapd, are restricted to work

only over USB. Additionally, the com.apple.file_relay service requires an Apple-signed
configuration profile to be installed. In iOS 11 or later, Apple TV can use the Secure Remote
Password protocol to wirelessly establish a pairing relationship.
A user can clear the list of trusted hosts with the Reset Network Settings or Reset Location
& Privacy options.
Mobile device management

Mobile device management security overview
Overview
Apple operating systems support mobile device management (MDM), which allows
organizations to securely configure and manage scaled Apple device deployments. MDM
capabilities are built on existing operating system technologies, such as configuration
profiles, over-the-air enrollment, and the Apple Push Notification service (APNs). For
example, APNs is used to wake the device so it can communicate directly with its MDM
solution over a secured connection. With APNs, no confidential or proprietary information
is transmitted.
Using MDM, IT departments can enroll Apple devices in an enterprise environment,

wirelessly configure and update settings, monitor compliance with corporate policies,
manage software update policies, and even remotely wipe or lock managed devices.
In addition to the traditional device enrollments supported by iOS, iPadOS, macOS, and
tvOS, an enrollment type has been added in iOS 13 or later, iPadOS 13.1 or later, and
macOS 10.15 or later—User Enrollment. User enrollments are MDM enrollments specifically
targeting “bring your own device” (BYOD) deployments where the device is personally
owned but used in a managed environment. User enrollments grant the MDM solution more
limited privileges than unsupervised device enrollments do, and provide cryptographic
separation of user and corporate data.

Enrollment types
• User Enrollment: User Enrollment is designed for devices owned by the user and
is integrated with Managed Apple IDs to establish a user identity on the device.
Managed Apple IDs are part of the User Enrollment profile, and the user must
successfully authenticate in order for enrollment to be completed. Managed Apple IDs
can be used alongside a personal Apple ID that the user has already signed in with.
Managed apps and accounts use a Managed Apple ID, and personal apps and accounts
use a personal Apple ID.
• Device Enrollment: Device Enrollment allows organizations to have users manually enroll
devices and then manage many different aspects of device use, including the ability to
erase the device. Device Enrollment also has a larger set of payloads and restrictions
that can be applied to the device. When a user removes an enrollment profile, all
configuration profiles, their settings, and managed apps based on that enrollment
profile are removed with it.
• Automated Device Enrollment: Automated Device Enrollment lets organizations

configure and manage devices from the moment the devices are removed from the
box (in a process known as Auto Advance deployment). These devices are known as
supervised, and users have the option to prevent the MDM profile from being removed
by the user. Automated Device Enrollment is designed for devices owned by the
organization.
Device restrictions
Restrictions can be enabled—or in some cases, disabled—by administrators to prevent
users from accessing a specific app, service, or function of an iPhone, iPad, Mac, or
Apple TV that’s enrolled in an MDM solution. Restrictions are sent to devices in a
restrictions payload, which is part of a configuration profile. Certain restrictions on an
iPhone may be mirrored on a paired Apple Watch.
Passcode and password settings management

By default, the user’s passcode can be defined as a numeric PIN. In iOS and iPadOS
devices with Touch ID or Face ID, the minimum passcode length is four digits. Because
longer and more complex passcodes are harder to guess or attack, they are recommended.
Administrators can enforce complex passcode requirements and other policies using MDM
or Microsoft Exchange ActiveSync, or by requiring users to manually install configuration
profiles. An administrator password is needed for the macOS passcode policy payload
installation. Some passcode policies can require a certain passcode length, composition, or
other attributes.

Configuration profile enforcement
Configuration profiles are the primary way that an MDM solution delivers and manages
policies and restrictions on managed devices. If organizations need to configure a large
number of devices—or to provide lots of custom email settings, network settings, or
certificates to a large number of devices—configuration profiles are a safe and secure way
to do it.
Configuration profiles
A configuration profile is an XML file (ending in .mobileconfig) that consists of payloads
that load settings and authorization information onto Apple devices. Configuration profiles
automate the configuration of settings, accounts, restrictions, and credentials. These
files can be created by an MDM solution or Apple Configurator 2, or they can be created
manually. Before organizations send a configuration profile to an Apple device, they must
enroll the device in the MDM solution using an enrollment profile.
Enrollment profiles
An enrollment profile is a configuration profile with an MDM payload that enrolls the device
in the MDM solution specified for that device. This allows the MDM solution to send
commands and configuration profiles to the device and to query certain aspects of the
device. When a user removes an enrollment profile, all configuration profiles, their settings,
and managed apps based on that enrollment profile are removed with it. There can be only
one enrollment profile on a device at a time.
Configuration profile settings

A configuration profile contains a number of settings in specific payloads that can be
specified, including (but not limited to):
• Passcode and password policies
• Restrictions on device features (for example, disabling the camera)
• Network and VPN settings
• Microsoft Exchange settings

• Mail settings
• Account settings
• LDAP directory service settings
• CalDAV calendar service settings
• Credentials and keys
• Software updates
Profile signing and encryption

Configuration profiles can be signed to validate their origin and encrypted to ensure
their integrity and protect their contents. Configuration profiles for iOS and iPadOS
are encrypted using the Cryptographic Message Syntax (CMS) specified in RFC 5652,
supporting 3DES and AES128.

Profile installation
Users can install configuration profiles directly on their devices using Apple Configurator 2,
or they can be downloaded using Safari, sent attached to a mail message, transferred
using AirDrop or the Files app in iOS and iPadOS, or sent over the air using a mobile device
management (MDM) solution. When a user sets up a device in Apple School Manager or
Apple Business Manager, the device downloads and installs a profile for MDM enrollment.
For information on how to remove profiles, see MDM Overview in MDM Settings for IT
Administrators.
Note: On supervised devices, configuration profiles can also be locked to a device to

completely prevent their removal or to allow removal only with a passcode. Because many
organizations own their iOS and iPadOS devices, configuration profiles that bind a device
to an MDM solution can be removed—but doing so also removes all managed configuration
information, data, and apps.
Automated Device Enrollment

Organizations can automatically enroll iOS, iPadOS, macOS, and tvOS devices in mobile
device management (MDM) without having to physically touch or prepare the devices
before users get them. After enrolling in one of the services, administrators sign in to the
service website and link the program to their MDM solution. The devices they purchased
can then be assigned to users through MDM. During the device configuration process,
security of sensitive data can be increased by ensuring appropriate security measures are
in place. For example:
• Have users authenticate as part of the initial setup flow in the Apple device’s Setup
Assistant during activation.
• Provide a preliminary configuration with limited access and require additional device
configuration to access sensitive data.
After a user has been assigned, any MDM-specified configurations, restrictions, or controls
are automatically installed. All communications between devices and Apple servers are
encrypted in transit through HTTPS (TLS).
The setup process for users can be further simplified by removing specific steps in the
Setup Assistant for devices, so users are up and running quickly. Administrators can also
control whether or not the user can remove the MDM profile from the device and ensure
that device restrictions are in place throughout the lifecycle of the device. After the
device is unboxed and activated, it can enroll in the organization’s MDM solution—and all
management settings, apps, and books are installed as defined by the MDM administrator.
Apple School Manager and Apple Business Manager

Apple School Manager and Apple Business Manager are services for IT administrators to
deploy Apple devices that an organization has purchased directly from Apple or through
participating Apple Authorized Resellers and carriers.

When used with an MDM solution, administrators can simplify the setup process for
users, configure device settings, and distribute apps and books purchased in Apple
School Manager and Apple Business Manager. Apple School Manager also integrates with
Student Information Systems (SISs) directly or using SFTP, and Apple School Manager
and Apple Business Manager can use System for Cross-domain Identity Management
(SCIM) or federated authentication with Microsoft Azure Active Directory (Azure AD) so
administrators can quickly create accounts.
Devices with iOS 11 or later and tvOS 10.2 or later can also be added to Apple
School Manager and Apple Business Manager after the time of purchase using
Apple Configurator 2.
Apple maintains certifications in compliance with the ISO/IEC 27001 and 27018 standards
to enable Apple customers to address their regulatory and contractual obligations.
These certifications provide our customers with an independent attestation over Apple’s
Information Security and Privacy practices for in-scope systems. For more information, see
the Apple Support article Apple Internet Services Certifications.
Note: To learn whether an Apple program is available in a specific country or region, see
the Apple Support article Availability of Apple programs for education and business.
Device supervision
Supervision generally denotes that the device is owned by the organization, giving them
additional control over the device’s configuration and restrictions.
iPhone and iPad devices with iOS 5 or later and Apple TV devices with tvOS 10.2 or later
become supervised by:
• Using Apple Configurator 2 to supervise the device
During this process, the device is erased and all data is lost.
• Enrolling the device in an MDM solution and selecting supervision as part of the
enrollment process
Mac computers can be supervised if they:
• Are running macOS 11 enrolled in MDM using device enrollment
• Are upgraded to macOS 11 and the enrollment in MDM was a user approved MDM
enrollment
• Are running macOS 10.14.4 or later and:
• The devices’ serial numbers appear in Apple School Manager or Apple Business
Manager
• Are enrolled in an MDM solution using Apple School Manager or Apple Business
Manager
The following devices are supervised automatically when enrolled in Apple School Manager
or Apple Business Manager:
• iPhone and iPod touch with iOS 13 or later
• iPad with iPadOS 13.1 or later
• Apple TV with tvOS 13 or later
• Mac computers with macOS 10.14.4 or later

Important: If the user knows the passcode, iPhone and iPad devices that aren’t supervised
can have manually installed configuration profiles removed, even if the option is set
to “never.” Manually installed configuration profiles for Mac computers can be removed
using the profiles command-line tool, or System Preferences if the user knows an
administrator’s user name and password. As of macOS 10.15, like on iOS and iPadOS,
profiles installed with MDM must be removed with MDM, or they are removed automatically
upon unenrollment from MDM.
Activation Lock security

How Apple enforces Activation Lock varies depending on whether the device is an iPhone
or an iPad, a Mac with Apple silicon, or an Intel-based Mac with the Apple T2 Security Chip.
Behavior on iPhone and iPad

On iPhone and iPad devices, Activation Lock is enforced through the activation process
after the Wi-Fi selection screen in iOS and iPadOS Setup Assistant. When device indicates
that its activating, it sends a request to an Apple server to get an activation certificate.
Devices that are Activation Locked prompt the user for the iCloud credentials of the user
that enabled Activation Lock at this time. iOS and iPadOS Setup Assistant won’t progress
unless a valid certificate can be obtained.
Behavior on a Mac with Apple silicon

In a Mac with Apple silicon, LLB verifies that a valid LocalPolicy for the device exists and
that the LocalPolicy policy nonce values match the values stored in the Secure Storage
Component. LLB boots to recoveryOS if:
• There is no LocalPolicy for the current macOS
• The LocalPolicy is invalid for that macOS
• The LocalPolicy nonce hash values don’t match the hashes of values stored in the
Secure Storage Component
recoveryOS detects that the Mac computer isn’t activated and contacts the activation
server to get an activation certificate. If the device is Activation Locked, recoveryOS
prompts the user for iCloud credentials of the user that enabled Activation Lock at this
time. After a valid activation certificate is obtained, that activation certificate key is used
to obtain a RemotePolicy certificate. The Mac computer uses the LocalPolicy key and
RemotePolicy certificate to produce a valid LocalPolicy. LLB won’t allow booting of macOS
unless a valid LocalPolicy is present.
Behavior on Intel-based Mac computers

In an Intel-based Mac with a T2 chip, the T2 chip firmware verifies that a valid activation
certificate is present before allowing the computer to boot to macOS. UEFI firmware loaded
by the T2 chip is responsible for querying the activation status of the device from the T2
chip and booting to recoveryOS instead of booting to macOS if a valid activation certificate
isn’t present. recoveryOS detects that the Mac isn’t activated and contacts the activation
server to get an activation certificate. If the device is Activation Locked, recoveryOS
prompts the user for iCloud credentials of the user that enabled Activation Lock at this
time. UEFI firmware won’t allow booting of macOS unless a valid activation certificate is
present.

Lost Mode, remote lock, and remote wipe
Managed Lost Mode is used to locate supervised devices when they are stolen. After they
are located, they can be remotely locked or erased.
Lost Mode
If a supervised iOS or iPadOS device with iOS 9 or later is lost or stolen, an MDM
administrator can remotely enable Lost Mode on that device. When Lost Mode is enabled,
the current user is logged out and the device can’t be unlocked. The screen displays a
message that can be customized by the administrator, such as displaying a phone number
to call if the device is found. When the device is put into Lost Mode, the administrator
can request the device to send its current location (even if Location Services are off) and,
optionally, play a sound. When an administrator turns off Lost Mode, which is the only way
the mode can be exited, the user is informed of this action through a message on the Lock
Screen or an alert on the Home screen.
Remote wipe, and remote lock

iOS, iPadOS, and macOS devices can be erased remotely by an administrator or user
(instant remote wipe is available only if the Mac has FileVault enabled). Instant remote wipe
is achieved by securely discarding the media key from Effaceable Storage, rendering all
data unreadable. A remote wipe command can be initiated by mobile device management
(MDM), Microsoft Exchange ActiveSync, or iCloud. On a Mac, the computer sends an
acknowledgment and performs the wipe. With a remote lock, MDM requires that a six-digit
passcode be applied to the Mac, rendering any user locked out until this passcode is typed
in.
Note: Remote lock isn’t available for a Mac with Apple silicon.
When a remote wipe command is triggered by MDM or iCloud, the device sends an
acknowledgment and performs the wipe. For remote wipe through Microsoft Exchange
ActiveSync, the device checks in with the Microsoft Exchange Server before performing
the wipe. Remote wipe isn’t possible in the following situations:
• With User Enrollment
• Using Microsoft Exchange ActiveSync when the account that was installed with User
Enrollment
• Using Microsoft Exchange ActiveSync if the device is supervised
Users can also wipe iOS and iPadOS devices in their possession using the Settings
app. And as mentioned, devices can be set to automatically wipe after a series of failed
passcode attempts.

Shared iPad security in iPadOS
Overview
Shared iPad is a multiuser mode for use in iPad deployments. It allows users to share an
iPad while maintaining separation of documents and data for each user. Each user gets
their own private, reserved storage location, which is implemented as an APFS (Apple
File System) volume protected by the user’s credential. Shared iPad requires the use of a
Managed Apple ID that’s issued and owned by the organization and enables a user to sign
in to any organizationally owned device that is configured for use by multiple users. User
data is partitioned into separate directories, each in their own data protection domains and
protected by both UNIX permissions and sandboxing. In iPadOS 13.4 or later, users can
also sign in to a temporary session. When the user signs out of a temporary session, their
APFS volume is deleted and its reserved space is returned to the system.
Signing in to Shared iPad

Both native and federated Managed Apple IDs are supported when signing in to
Shared iPad. When using a federated account for the first time, the user is redirected to
the Identity Provider’s (IdP) sign-in portal. After authenticated, a short-lived access token
is issued for the backing Managed Apple IDs—and the login process proceeds similarly
to the native Managed Apple IDs sign-in process. Once signed in, Setup Assistant on
Shared iPad prompts the user to establish a passcode (credential) used to secure the local
data on the device and to authenticate to the login screen in the future. Like a single-
user device, in which the user would sign in once to their Managed Apple ID using their
federated account and then unlock their device with their passcode, on Shared iPad the
user signs in once using their federated account and from then on uses their established
passcode.
When a user signs in without federated authentication, the Managed Apple ID is

authenticated with Apple Identity Service (IDS) using the SRP protocol. If authentication
is successful, a short-lived access token specific to the device is granted. If the user has
used the device before, they already have a local user account, which is unlocked using
the same credential.
If the user hasn’t used the device before or is using the temporary session feature,
Shared iPad provisions a new UNIX user ID, an APFS volume to store the user’s personal
data, and a local keychain. Because storage is allocated (reserved) for the user at the time
the APFS volume is created, there may be insufficient space to create a new volume. In
such an event, the system will identify an existing user whose data has finished syncing to
the cloud and evict that user from the device in order to allow the new user to sign in. In
the unlikely event that all existing users haven’t completed uploading their cloud data, the
new user sign in fails. To sign in, the new user will need to wait for one user’s data to finish
syncing, or have an administrator forcibly delete an existing user account, thereby risking
data loss.
If the device isn’t connected to the internet (for example, if the user has no Wi-Fi access
point), authentication can occur against the local account for a limited number of days. In
that situation, only users with previously existing local accounts or a temporary session
can sign in. After the time limit has expired, users are required to authenticate online, even
if a local account already exists.

After a user’s local account has been unlocked or created, if it’s remotely authenticated,
the short-lived token issued by Apple’s servers is converted to an iCloud token that
permits signing in to iCloud. Next, the users’ settings are restored and their documents
and data are synced from iCloud.
While a user session is active and the device remains online, documents and data are
stored on iCloud as they are created or modified. In addition, a background syncing
mechanism ensures that changes are pushed to iCloud, or to other web services using
NSURLSession background sessions, after the user signs out. After background syncing for
that user is complete, the user’s APFS volume is unmounted and can’t be mounted again
without the user signing back in.
Temporary sessions don’t sync data with iCloud, and although a temporary session can
sign into a third-party syncing service such as Box or Google Drive, there’s no facility to
continue syncing data when the temporary session ends.
Signing out of Shared iPad

When a user signs out of Shared iPad, that user’s keybag is immediately locked and all
apps are shut down. To accelerate the case of a new user signing in, iPadOS defers some
ordinary sign-out actions temporarily and presents a login window to the new user. If a user
signs in during this time (approximately 30 seconds), Shared iPad performs the deferred
cleanup as part of signing in to the new user account. However, if Shared iPad remains idle,
it triggers the deferred cleanup. During the cleanup phase, Login Window is restarted as if
another sign-out had occurred.
When a temporary session is ended, Shared iPad performs the full logout sequence and
deletes the temporary session’s APFS volume immediately.

Apple Configurator 2 security
Apple Configurator 2 features a flexible, secure, device-centric design that enables an
administrator to quickly and easily configure one or dozens of iOS, iPadOS, and tvOS
devices connected to a Mac through USB (or tvOS devices paired through Bonjour) before
giving them to users. With Apple Configurator 2, an administrator can update software,
install apps and configuration profiles, rename and change wallpaper on devices, export
device information and documents, and much more.
Administrators can also choose to add iOS, iPadOS, and tvOS devices to Apple School
Manager or Apple Business Manager using Apple Configurator 2, even if the devices
weren’t purchased directly from Apple, an Apple Authorized Reseller, or an authorized
cellular carrier. When the administrator sets up a device that has been manually enrolled,
it behaves like any other enrolled device, with mandatory supervision and mobile device
management (MDM) enrollment. For devices that weren’t purchased directly, the user has
a 30-day provisional period to remove the device from enrollment, supervision, and MDM.
The 30-day provisional period begins after the device is activated.
If iOS, iPadOS, and tvOS devices that have absolutely no internet connection and that are
connected to a host Mac with an internet connection while the devices are being set up,
organizations can use Apple Configurator 2 to activate them. Administrators can restore,
activate, and prepare devices with their necessary configuration including Apps, Profiles,
and Documents without ever needing to connect to either Wi-Fi or cellular networks. This
feature doesn’t allow an administrator to bypass any existing Activation Lock requirements
normally required during nontethered activation.

Screen Time security
Screen Time is a feature—in iOS 12 or later, iPadOS, and macOS 10.15 or later, and some
features of watchOS 6 or later—that lets users understand and control their own app and
web usage, or that of their children. Although Screen Time isn’t a new system security
feature, it’s important to understand how Screen Time protects the security and privacy of
the data gathered and shared between devices.
In Screen Time, there are two types of users: adults and children.
The table below describes the main features of Screen Time.
Feature Supported operating system
View usage data iO S
iPadO S
macO S
Enforce additional restrictions iO S
iPadO S
macO S
watchO S
Set web usage limits iO S
iPadO S
macO S
Set app limits iO S
iPadO S
macO S
watchO S
Configure Downtime iO S
iPadO S
macO S
watchO S
For users managing their own device usage, Screen Time controls and usage data can be
synced across devices associated to the same iCloud account using CloudKit end-to-end
encryption. This requires that the user’s account have two-factor authentication enabled
(syncing is on by default). Screen Time replaces the Restrictions feature found in previous
versions of iOS and iPadOS, and the Parental Controls feature found in previous versions of
macOS.
In iOS 13 or later, iPadOS 13.1 or later, and macOS 10.15 or later, Screen Time users and
managed children automatically share their usage across devices if their iCloud account
has two-factor authentication enabled. When a user clears Safari history or deletes an app,
the corresponding usage data is removed from the device and all synced devices.

Parents and Screen Time
Parents can also use Screen Time in iOS, iPadOS, and macOS devices to understand and
control their children’s use. If the parent is a family organizer (in iCloud Family Sharing),
they can view usage data and manage Screen Time settings for their children. Children are
informed when their parents turn on Screen Time, and they can monitor their own usage
as well. When parents turn on Screen Time for their children, the parents set a passcode
so their children can’t make changes. When they reach age of majority (age will vary
depending on country or region), the children can turn this monitoring off.
Usage data and configuration settings are transferred between the parentʼs and childʼs
devices using the end-to-end encrypted Apple Identity Service (IDS) protocol. Encrypted
data may be briefly stored on IDS servers until it’s read by the receiving device (for
example, as soon as the iPhone, iPad, or iPod touch is turned on, if it was off). This data
isn’t readable by Apple.
Screen Time analytics

If the user turns on Share iPhone & Watch Analytics, only the following anonymized data is
collected so that Apple can better understand how Screen Time is being used:
• Was Screen Time turned on during Setup Assistant or later in Settings
• Change in Category usage after creating a limit for it (within 90 days)
• Is Screen Time turned on
• Is Downtime enabled
• Number of times the “Ask for more” query was used
• Number of app limits
• Number of times users viewed usage in the Screen Time settings, per user type and per
view type (local, remote, widget)
• Number of times users ignore a limit, per user type
• Number of times users delete a limit, per user type
No specific app or web usage data is gathered by Apple. When a user sees a list of apps in
Screen Time usage information, the app icons are pulled directly from the App Store, which
doesn’t retain any data from these requests.

Glossary
Address Space Layout Randomization (ASLR) A technique employed by operating

systems to make the successful exploitation by a software bug much more difficult. By
ensuring memory addresses and offsets are unpredictable, exploit code can’t hard code
these values.
AES (Advanced Encryption Standard) A popular global encryption standard used to

encrypt data to keep it private.
AES cryptographic engine A dedicated hardware component that implements AES.
AES-XTS A mode of AES defined in IEEE 1619-2007 meant to work for encrypting storage
media.
APFS (Apple File System) The default file system for iOS, iPadOS, tvOS, watchOS, and
Mac computers using macOS 10.13 or later. APFS features strong encryption, space
sharing, snapshots, fast directory sizing, and improved file system fundamentals.
Apple Business Manager Apple Business Manager is a simple, web-based portal for IT
administrators that provides a fast, streamlined way for organizations to deploy Apple
devices that they have purchased directly from Apple or from a participating Apple
Authorized Reseller or carrier. They can automatically enroll devices in their mobile device
management (MDM) solution without having to physically touch or prepare the devices
before users get them.
Apple Identity Service (IDS) Apple’s directory of iMessage public keys, APNs addresses,
and phone numbers and email addresses that are used to look up the keys and device
addresses.
Apple Push Notification service (APNs) A worldwide service provided by Apple that
delivers push notifications to Apple devices.
Apple School Manager Apple School Manager is a simple, web-based portal for IT
administrators that provides a fast, streamlined way for organizations to deploy Apple
devices that they have purchased directly from Apple or from a participating Apple
Authorized Reseller or carrier. They can automatically enroll devices in their mobile device
management (MDM) solution without having to physically touch or prepare the devices
before users get them.
Apple Security Bounty A reward given by Apple to researchers who report a vulnerability
that affects the latest shipping operating systems and, where relevant, the latest hardware.
Boot Camp Boot Camp supports the installation of Microsoft Windows on supported Mac
computers.

Boot Progress Register (BPR) A set of system on chip (SoC) hardware flags that software
can use to track the boot modes the device has entered, such as Device Firmware Update
(DFU) mode and Recovery mode. After a Boot Progress Register flag is set, it can’t be
cleared. This allows later software to get a trusted indicator of the state of the system.
Boot ROM The very first code executed by a device’s processor when it first boots. As an
integral part of the processor, it can’t be altered by either Apple or an attacker.
CKRecord A dictionary of key-value pairs that contain data saved to or fetched from
CloudKit.
Data Protection File and Keychain protection mechanism for supported Apple devices. It
can also refer to the APIs that apps use to protect files and keychain items.
Data Vault A mechanism—enforced by the kernel—to protect against unauthorized access

to data regardless of whether the requesting app is itself sandboxed.
Device Firmware Upgrade (DFU) mode A mode in which a device’s Boot ROM code waits
to be recovered over USB. The screen is black when in DFU mode, but upon connecting
to a computer running iTunes or the Finder, the following prompt is presented: “iTunes (or
the Finder) has detected an (iPad, iPhone, or iPod touch) in Recovery mode. The user must
restore this (iPad, iPhone, or iPod touch) before it can be used with iTunes (or the Finder).”
direct memory access (DMA) A feature that enables hardware subsystems to access main
memory independent of the CPU.
Effaceable Storage A dedicated area of NAND storage, used to store cryptographic keys,
that can be addressed directly and wiped securely. While it doesn’t provide protection if an
attacker has physical possession of a device, keys held in Effaceable Storage can be used
as part of a key hierarchy to facilitate fast wipe and forward security.
Elliptic Curve Diffie-Hellman Exchange (ECDHE) Elliptic Curve Diffie-Hellman Exchange

with ephemeral keys. ECDHE allows two parties to agree on a secret key in a way that
prevents the key from being discovered by an eavesdropper watching the messages
between the two parties.
Elliptic Curve Digital Signature Algorithm (ECDSA) Elliptic Curve Digital Signature
Algorithm (ECDSA) is a digital signature algorithm based on elliptic curve cryptography.
eSPI The Enhanced Serial Peripheral Interface bus for synchronous serial communication.
Exclusive Chip Identification (ECID) A 64-bit identifier that’s unique to the processor
in each iOS and iPadOS device. When a call is answered on one device, ringing of nearby
iCloud-paired devices is terminated by briefly advertising through Bluetooth Low Energy
(BLE) 4.0. The advertising bytes are encrypted using the same method as Handoff
advertisements. Used as part of the personalization process, it’s not considered a secret.
file system key The key that encrypts each file’s metadata, including its class key. This is
kept in Effaceable Storage to facilitate fast wipe, rather than confidentiality.
group ID (GID) Like the UID, but common to every processor in a class.
hardware security module (HSM) A specialized tamper-resistant computer that

safeguards and manages digital keys.
HMAC A hash-based message authentication code based on a cryptographic hash

function.

iBoot Code that loads XNU, as part of the secure boot chain. Depending on the system on
chip (SoC) generation, iBoot may be loaded by the Low-Level Bootloader or directly by the
Boot ROM.
Input/Output Memory Management Unit (IOMMU) An input/output memory management

unit. A subsystem in an integrated chip that controls access to address space from other
input/output devices and peripherals.
integrated circuit (IC) Also known as a microchip.
Joint Test Action Group (JTAG) A standard hardware debugging tool used by
programmers and circuit developers.
keybag A data structure used to store a collection of class keys. Each type (user, device,
system, backup, escrow, or iCloud Backup) has the same format.
A header containing: Version (set to four in iOS 12 or later), Type (system, backup, escrow,
or iCloud Backup), Keybag UUID, an HMAC if the keybag is signed, and the method used
for wrapping the class keys—tangling with the UID or PBKDF2, along with the salt and
iteration count.
A list of class keys: Key UUID, Class (which file or Keychain Data Protection class),
wrapping type (UID-derived key only; UID-derived key and passcode-derived key),
wrapped class key, and a public key for asymmetric classes.
keychain The infrastructure and a set of APIs used by Apple operating systems and third-
party apps to store and retrieve passwords, keys, and other sensitive credentials.
key wrapping Encrypting one key with another. iOS and iPadOS use NIST AES key
wrapping, in accordance with RFC 3394.
Low-Level Bootloader (LLB) On Mac computers with a two-stage boot architecture, LLB
contains the code that’s invoked by the Boot ROM and that in turn loads iBoot, as part of
the secure boot chain.
media key Part of the encryption key hierarchy that helps provide for a secure and instant
wipe. In iOS, iPadOS, tvOS, and watchOS, the media key wraps the metadata on the
data volume (and thus without it access to all per-file keys is impossible, rendering files
protected with Data Protection inaccessible). In macOS, the media key wraps the keying
material, all metadata, and data on the FileVault protected volume. In either case, wipe of
the media key renders encrypted data inaccessible.
memory controller The subsystem in a system on chip that controls the interface between
the system on chip and its main memory.
mobile device management (MDM) A service that lets the user remotely manage enrolled
devices. After a device is enrolled, the user can use the MDM service over the network to
configure settings and perform other tasks on the device without user interaction.
NAND Nonvolatile flash memory.
per-file key The key used by Data Protection to encrypt a file on the file system. The per-
file key is wrapped by a class key and is stored in the fileʼs metadata.

provisioning profile A property list (.plist file) signed by Apple that contains a set of
entities and entitlements allowing apps to be installed and tested on an iOS or iPadOS
device. A development provisioning profile lists the devices that a developer has chosen
for ad hoc distribution, and a distribution provisioning profile contains the app ID of an
enterprise-developed app.
Recovery mode A mode used to restore many Apple devices if it doesn’t recognize the
user’s device so the user can reinstall the operating system.
ridge flow angle mapping A mathematical representation of the direction and width of the
ridges extracted from a portion of a fingerprint.
Secure Storage Component On supported devices, the Secure Enclave is paired with a
Secure Storage Component for anti-replay nonce storage. A chip designed with immutable
RO code, a hardware random number generator, cryptography engines, and physical
tamper detection. To read and update nonces, the Secure Enclave and storage chip
employ a secure protocol that ensures exclusive access to the nonces. There are multiple
generations of this technology with differing security guarantees.
sepOS The Secure Enclave firmware, based on an Apple-customized version of the L4

microkernel.
software seed bits Dedicated bits in the Secure Enclave AES Engine that get appended to
the UID when generating keys from the UID. Each software seed bit has a corresponding
lock bit. The Secure Enclave Boot ROM and operating system can independently change
the value of each software seed bit as long as the corresponding lock bit hasn’t been set.
After the lock bit is set, neither the software seed bit nor the lock bit can be modified. The
software seed bits and their locks are reset when the Secure Enclave reboots.
SSD controller A hardware subsystem that manages the storage media (solid-state drive).
System Coprocessor Integrity Protection (SCIP) A mechanism Apple uses to prevent

modification of coprocessor firmware.
system on chip (SoC) An integrated circuit (IC) that incorporates multiple components
into a single chip. The Application Processor, the Secure Enclave, and other coprocessors
are components of the SoC.
system software authorization A process that combines cryptographic keys built

into hardware with an online service to check that only legitimate software from Apple,
appropriate to supported devices, is supplied and installed at upgrade time.
tangling The process by which a user’s passcode is turned into a cryptographic key and
strengthened with the device’s UID. This process ensures that a brute-force attack must
be performed on a given device, and thus is rate limited and can’t be performed in parallel.
The tangling algorithm is PBKDF2, which uses AES keyed with the device UID as the
pseudorandom function (PRF) for each iteration.
UEFI firmware Unified Extensible Firmware Interface, a replacement technology for BIOS
to connect firmware to a computer’s operating system.
Uniform Resource Identifier (URI) A string of characters that identifies a web-based

resource.

unique ID (UID) A 256-bit AES key that’s burned into each processor at manufacture. It
can’t be read by firmware or software, and it’s used only by the processor’s hardware AES
Engine. To obtain the actual key, an attacker would have to mount a highly sophisticated
and expensive physical attack against the processor’s silicon. The UID isn’t related to any
other identifier on the device including, but not limited to, the UDID.
xART eXtended Anti-Replay Technology or a set of services that provides encrypted,

authenticated persistent storage for the Secure Enclave with anti-replay capabilities based
on the physical storage architecture. See Secure Storage Component.
XNU The kernel at the heart of the Apple operating systems. It’s assumed to be trusted,
and it enforces security measures such as code signing, sandboxing, entitlement checking,
and Address Space Layout Randomization (ASLR).

Document revision history
Date Summary
February 2021 Updated for:
• iO S 14.3
• iPadO S 14.3
• macO S 11.1
• tvO S 14.3
• watchO S 7.2
Topics added:
• Memory safe iB oot implementation

• Boot process for a Mac with Apple silicon
• Boot modes for a Mac with Apple silicon
• Startup Disk security policy control for a Mac with
Apple silicon
• LocalPolicy signing-key creation and management
• Contents of a LocalPolicy file for a Mac with Apple
silicon
• Signed system volume security in macO S
• Apple Security Research Device
• Password Monitoring
• IPv6 security
• Car keys security in iO S
Topics updated:
• Secure Enclave
• Hardware microphone disconnect
• recoveryO S and diagnostics environments for an
Intel-based Mac
• Direct memory access protections for Mac
computers
• Kernel extensions in macO S
• System Integrity Protection
• System security for watchO S
• Managing FileVault in macO S
• App access to saved passwords
• Password security recommendations
• Apple Cash security in iO S, iPadO S, and watchOS
• Secure Business Chat using the Messages app
• Wi-Fi privacy
• Activation Lock security
• Apple Configurator 2 security

Date Summary
April 2020 Updated for:
• iO S 13.4
• iPadO S 13.4
• macO S 10.15.4
• tvO S 13.4
• watchO S 6.2
Updates:
• iPad microphone disconnect added to Hardware

microphone disconnect.
• Data vaults added to Protecting app access to user
data.
• Updates to Managing FileVault in macO S and
Command-line tools.
• Malware Removal Tool additions in Protecting
against malware in macO S.
• Updates to Shared iPad security in iPadO S.
December 2019 Merged the iO S Security Guide, macO S Security

Overview, and the Apple T2 Security Chip Overview
Updated for:
• iO S 13.3
• iPadO S 13.3
• macO S 10.15.2
• tvO S 13.3
• watchO S 6.1.1
Privacy Controls, Siri and Siri Suggestions, and Safari
Intelligent Tracking Prevention have been removed. See
https://www.apple.com/privacy/ for the latest on those
features.
May 2019 Updated for iO S 12.3
• Support for TLS 1.3

• Revised description of AirDrop security
• DFU mode and Recovery mode
• Passcode requirements for accessory connections
November 2018 Updated for iO S 12.1
• Group FaceTime
September 2018 Updated for iO S 12
• Secure Enclave
• OS Integrity Protection
• Express Card with power reserve
• DFU mode and Recovery mode
• HomeKit TV Remote accessories
• Contactless passes
• Student ID cards
• Siri Suggestions
• Shortcuts in Siri
• Shortcuts app
• User password management
• Screen Time
• Security Certifications and programs

Date Summary
July 2018 Updated for iO S 11.4
• Biometric policies
• HomeKit
• Apple Pay
• Business Chat
• Messages in iCloud
• Apple Business Manager
December 2017 Updated for iO S 11.2
• Apple Pay Cash
October 2017 Updated for iO S 11.1

• Touch ID/Face ID
• Shared Notes
• CloudKit end-to-end encryption
• TLS update
• Apple Pay, Paying with Apple Pay on the web
• Siri Suggestions
• Shared iPad
July 2017 Updated for iO S 10.3
• Secure Enclave
• File Data Protection
• Keybags
• SiriKit
• HealthKit
• Network Security
• Bluetooth
• Shared iPad
• Lost Mode
• Activation Lock
• Privacy Controls
March 2017 Updated for iO S 10
• System Security
• Data Protection classes
• HomeKit, ReplayKit, SiriKit
• Apple Watch
• Wi-Fi, VPN
• Single sign-on
• Apple Pay, Paying with Apple Pay on the web
• Credit, debit, and prepaid card provisioning
• Safari Suggestions

Date Summary
May 2016 Updated for iO S 9.3
• Managed Apple ID
• Two-factor authentication for Apple ID
• Keybags
• Security Certifications
• Lost Mode, Activation Lock
• Secure Notes
• Apple School Manager
• Shared iPad
September 2015 Updated for iO S 9
• Apple Watch Activation Lock

• Passcode policies
• Touch ID API support
• Data Protection on A8 uses AES-XTS
• Keybags for unattended software update
• Certification updates
• Enterprise app trust model
• Data Protection for Safari bookmarks
• App Transport Security
• VPN specifications
• iCloud Remote Access for HomeKit
• Apple Pay Rewards cards, Apple Pay card issuer’s
app
• Spotlight on-device indexing
• iO S Pairing Model
• Apple Configurator 2
• Restrictions

Apple Inc.
© 2021 Apple Inc. All rights reserved.
Use of the “keyboard” Apple logo (Option-Shift-K) for commercial purposes without the prior written consent of
Apple may constitute trademark infringement and unfair competition in violation of federal and state laws.
Apple, the Apple logo, AirDrop, AirPlay, Apple CarPlay, Apple Music, Apple Pay, Apple TV, Apple Watch, CloudKit,
Face ID, FaceTime, FileVault, Finder, FireWire, Handoff, HomeKit, HomePod, iMac, iMac Pro, iMessage, iPad,
iPad Air, iPadOS, iPhone, iPod touch, iTunes, , Keychain, Lightning, Mac, MacBook, MacBook Air, MacBook Pro,
macOS, Mac Pro, Objective-C, OS X, QuickType, Safari, Siri, Siri Remote, Spotlight, Touch ID, TrueDepth, watchOS,
and Xcode are trademarks of Apple Inc., registered in the U.S. and other countries.
Apple Books, Apple Wallet, HealthKit, SiriKit, Touch Bar, and tvOS are trademarks of Apple Inc.
AppleCare, App Store, iCloud, iCloud Drive, iCloud Keychain, and iTunes Store are service marks of Apple Inc.,
registered in the U.S. and other countries.
IOS is a trademark or registered trademark of Cisco in the U.S. and other countries and is used under license.
The Bluetooth® word mark and logos are registered trademarks owned by Bluetooth SIG, Inc. and any use of such
marks by Apple is under license.
Java is a registered trademark of Oracle and/or its affiliates.
UNIX® is a registered trademark of The Open Group.
Other product and company names mentioned herein may be trademarks of their respective companies. Product
specifications are subject to change without notice.
Apple
One Apple Park Way
Cupertino, CA 95014
USA
apple.com
028-00309

Apple Platform Security Guide

Uploaded by

Copyright:

Available Formats

Apple Platform Security Guide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Apple Platform Security Guide

Uploaded by

Copyright:

Available Formats

Apple

Introduction to Apple platform security 5

Hardware security and biometrics 7

Hardware security overview 7

System security overview 25

Encryption and Data Protection 70

Encryption and Data Protection overview 70

Apple Platform Security 2

App security overview 95

Services security 107

Services security overview 107

Network security 149

Network security overview 149

Apple Platform Security 3

Developer kit security overview 163

Secure device management 174

Secure device management overview 174

Document revision history 192

Apple Platform Security 4

The content is organized into the following topic areas:

• Services security: Apple’s services for identification, password management, payments,

Apple Platform Security 5

• Secure device management: Methods that allow management of Apple devices,

Apple Platform Security 6

Hardware security overview

Apple Platform Security 7

Feature A10 A11, S3 A12, S4 A13, S5 A14, S6 M1

Page See Note

Feature A10 A11, S3 A12, S4 A13, S5 A14, M1, S6

Apple Platform Security 8

The Secure Enclave components.

Apple Platform Security 9

• iPad Air or later

• Intel-based Mac computers that contain the Apple T2 Security Chip

• Mac computers with Apple silicon

• Apple Watch Series 1 or later

• HomePod and HomePod mini

Secure Enclave Processor

Memory Protection Engine

Apple Platform Security 10

Secure Enclave Boot ROM

Apple Platform Security 11

At system startup, the Secure Enclave Processor’s System Coprocessor Integrity

True Random Number Generator

Root Cryptographic Keys

Apple Platform Security 12

Apple Platform Security 13

Secure nonvolatile storage

Apple Platform Security 14

• Enabling or disabling Touch ID or Face ID

• Adding or removing a Touch ID fingerprint or Face ID face

• Touch ID or Face ID reset

• Adding or removing an Apple Pay card

• Erase All Content and Settings

On architectures that don’t feature a Secure Storage Component, EEPROM (electrically

Apple Platform Security 15

Power and clock monitors

Secure Enclave feature summary

SoC Memory Protection Secure Storage AES Engine PKA