CISCO Wireless LAN Security
CISCO Wireless LAN Security
CISCO Wireless LAN Security
Authors
Sean Convery (CCIE #4232), Darrin Miller (CCIE #6447), and Sri Sundaralingam are
the primary authors of this white paper. Mark Doering, Pej Roshan, Stacey Albert,
Bruce McMurdo, and Jason Halpern provided significant contributions to this paper
and are the lead architects of Cisco’s reference implementation in San Jose,
California, USA. All are network architects who focus on wireless LAN, VPN, or
security issues.
Audience
Though this document is technical in nature, it can be read at different levels of detail, depending on your level of
interest. A network manager, for example, can read the introductory sections in each area to obtain a good overview
of security design strategies and consideration for WLAN networks. A network engineer or designer can read this
document in its entirety and gain design information and threat analysis details, which are supported by actual
configuration snapshots for the devices involved. Because this document covers a wide range of WLAN deployments,
it may be helpful to read the introductory sections of the paper first and then skip right to the type of WLAN you
are interested in deploying.
Caveats
This document presumes that you already have a security policy in place. Cisco Systems does not recommend
deploying WLANs—or any networking technology—without an associated security policy. Although network
security fundamentals are mentioned in this document, they are not described in detail. Security within this document
is always mentioned as it pertains to WLANs.
Even though WLANs introduce security risks, many organizations choose to deploy WLANs because they bring user
productivity gains and simplify deployment of small networks. Following the guidelines in this document does not
guarantee a secure WLAN environment, nor does it guarantee that you will prevent all penetrations. By following
the guidelines, you will mitigate WLAN security risks as much as possible.
Though this document contains a large amount of detail on most aspects of wireless security, the discussion is not
exhaustive. In particular, the document does not address wireless bridges, personal digital assistants (PDAs), or
non-802.11-based WLAN technology. In addition, it does not provide specific best practices on general WLAN
deployment and design issues that are not security related.
During the validation of SAFE, real products were configured in the exact network implementation described in this
document. Specific configuration snapshots from the lab are included in Appendix A, “Validation Lab.”
Throughout this document the term “hacker” denotes an individual who attempts to gain unauthorized access to
network resources with malicious intent. Although the term “cracker” is generally regarded as the more accurate
word for this type of individual, hacker is used here for readability.
Architecture Overview
Design Fundamentals
Cisco SAFE wireless emulates as closely as possible the functional requirements of today’s networks. Implementation
decisions varied, depending on the network functionality required. However, the following design objectives, listed
in order of priority, guided the decision-making process:
• Security and attack mitigation based on policy
• Authentication and authorization of users to wired network resources
First and foremost, SAFE wireless needs to provide a secure WLAN connectivity option to enterprise networks. As
a connectivity option, WLAN access must adhere to an organization’s security policy as closely as possible. In
addition, it must provide this access as securely as possible while recognizing the need to maintain as many of the
characteristics of a traditional wired LAN as possible. Finally, WLANs must integrate with existing network designs
based on the SAFE security architecture.
Wireless networks have become one of the most interesting targets for hackers today. Organizations today are
deploying wireless technology at a rapid rate, often without considering all security aspects. This rapid deployment
is due, in part, to the low cost of the devices, ease of deployment, and the large productivity gains. Because WLAN
devices ship with all security features disabled, increasing WLAN deployments have attracted the attention of the
hacker community. Several Web sites document freely available wireless connections throughout the United States.
Although most hackers are using these connections as a means to get free Internet access or to hide their identity, a
smaller group sees this situation as an opportunity to break into networks that otherwise might have been difficult
to attack from the Internet. Unlike a wired network, a WLAN sends data over the air and may be accessible outside
the physical boundary of an organization. When WLAN data is not encrypted, the packets can be viewed by anyone
within radio frequency range. For example, a person with a Linux laptop, a WLAN adapter, and a program such as
TCPDUMP can receive, view, and store all packets circulating on a given WLAN.
MAC Authentication
WLAN access points can identify every wireless card ever manufactured by its unique Media Access Control (MAC)
address that is burned into and printed on the card. Some WLANs require that the cards be registered before the
wireless services can be used. The access point then identifies the card by the user, but this scenario is complex because
every access point needs to have access to this list. Even if it were implemented, it cannot account for hackers who
use WLAN cards that can be loaded with firmware that does not use the built-in MAC address, but a randomly
chosen, or deliberately spoofed, address. Using this spoofed address, a hacker can attempt to inject network traffic
or spoof legitimate users.
The security impact of ad hoc WLANs is significant. Many wireless cards, including some shipped as a default item
by PC manufacturers, support ad hoc mode. When adapters use ad hoc mode, any hacker with an adapter configured
for ad hoc mode and using the same settings as the other adapters may gain unauthorized access to clients.
A rogue access point is one that is accessible to an organization’s employees but is not managed as a part of the trusted
network. Most rogue access points are installed by employees for which IT is not providing WLAN access. A typical
rogue access point, then, is an inexpensive one that an employee purchases and installs by plugging it into an
available switch port, often with no security measures enabled. A hacker, even one outside the physical boundaries
of an organization’s facilities, can gain access to the trusted network simply by associating with a rogue access point.
Another type of rogue access point is one that masquerades as a trusted access point and tricks WLAN users into
associating with it, thereby enabling a hacker to manipulate wireless frames as they cross the access point.
The threat posed by rogue access points can be mitigated by preventing their deployment and detecting those rogue
access points that are deployed. The following components are required in order to mitigate the threat of rogue access
points. A detailed discussion of these points can be found in Appendix C, “Rogue Access Point Additional
Information.”
Prevention
• Corporate policy
• Physical security
• Supported WLAN infrastructure
• 802.1X port-based security on edge switches
Detection
• Using wireless analyzers or sniffers
• Using scripted tools on the wired infrastructure
• Physically observing WLAN access point placement and usage
As discussed in the primer (Appendix B), 802.11b and 802.11a are the most widely deployed WLAN technologies
today. Traditional 802.11 WLAN security includes the use of open or shared-key authentication and static wired
equivalent privacy (WEP) keys. This combination offers a rudimentary level of access control and privacy, but each
element can be compromised. The following sections describe these elements and the challenges of their use in
enterprise environments.
Authentication
The 802.11 standard supports two means of client authentication: open and shared-key authentication. Open
authentication involves little more than supplying the correct service set ID (SSID). With open authentication, the use
of WEP prevents the client from sending data to and receiving data from the access point, unless the client has the
correct WEP key. With shared-key authentication, the access point sends the client device a challenge text packet that
the client must then encrypt with the correct WEP key and return to the access point. If the client has the wrong key
or no key, authentication will fail and the client will not be allowed to associate with the access point. Shared-key
authentication is not considered secure because a hacker who detects both the clear text challenge and the same
challenge encrypted with a WEP key can decipher the WEP key.
Key Management
Another type of key that is often used—but is not considered secure—is a “static” WEP key. A static WEP key is a
key composed of either 40 or 128 bits that is statically defined by the network administrator on the access point and
all clients that communicate with the access point. When static WEP keys are used, a network administrator must
perform the time-consuming task of entering the same keys on every device in the WLAN.
If a device that uses static WEP keys is lost or stolen, the possessor of the stolen device can access the WLAN. An
administrator will not be able to detect that an unauthorized user has infiltrated the WLAN until and unless the theft
is reported. The administrator must then change the WEP key on every device that uses the same static WEP key used
by the missing device. In a large enterprise WLAN with hundreds or even thousands of users, this can be a daunting
task. Worse still, if a static WEP key is deciphered through a tool such as AirSnort, the administrator has no way of
knowing that the key has been compromised by a hacker.
WEP
The 802.11 standards define WEP as a simple mechanism to protect the over-the-air transmission between WLAN
access points and network interface cards (NICs). Working at the data link layer, WEP requires that all
communicating parties share the same secret key. To avoid conflicting with U.S. export controls that were in effect
at the time the standard was developed, 40-bit encryption keys were required by IEEE 802.11b, though many
vendors now support the optional 128-bit standard. WEP can be easily cracked in both 40- and 128-bit variants by
using off-the-shelf tools readily available on the Internet. On a busy network, 128-bit static WEP keys can be
obtained in as little as 15 minutes, according to current estimates. These attacks are described in more detail in the
following paragraphs.
The initialization vector is at the center of most of the issues that involve WEP. Because the initialization vector is
transmitted as plaintext and placed in the 802.11 header, anyone sniffing a WLAN can see it. At 24 bits long, the
initialization vector provides a range of 16,777,216 possible values. A University of California, Berkeley paper found
that when the same initialization vector is used with the same key on an encrypted packet (known as an
initialization-vector collision), a hacker can capture the data frames and derive information about the data as well
as the network. For more information, refer to the paper at:
http://www.isaac.cs.berkeley.edu/isaac/wep-faq.html
In the past year, encryption analysts from the University of California, Berkeley, the University of Maryland, and
Cisco Systems, Inc. have reported weaknesses in the authentication and WEP encryption schemes in the IEEE 802.11
WLAN standard. These researchers have called for sophisticated key management solutions to mitigate these flaws.
The University of Maryland paper can be found at:
http://www.cs.umd.edu/~waa/wireless.pdf
Cryptanalysts Fluhrer, Mantin, and Shamir (FMS) discovered inherent shortcomings with the RC4 key-scheduling
algorithm. Because RC4 as implemented in WEP uses a 24-bit initialization vector and does not dynamically rotate
encryption keys, these shortcomings are demonstrated to have practical applications in decrypting 802.11 frames
using WEP. The attack illustrated in the paper focuses on a large class of weak initialization vectors that can be
generated by RC4, and highlights methods to break the key using certain patterns in the initialization vectors. This
attack is pragmatic, but the most disconcerting fact is that the attack is completely passive. In this paper, this attack
is known as the FMS attack. The FMS attack discusses the theoretical derivation of a WEP key in a range of 100,000
to 1,000,000 packets encrypted using the same key. More detail can be found in the paper itself at:
http://www.cs.umd.edu/~waa/class-pubs/rc4_ksaproc.ps
Recent practical implementations of the FMS attack have been able to derive a static WEP key by capturing about a
million packets. This is demonstrated in a paper by AT&T Labs and Rice University at the following URL:
http://www.cs.rice.edu/~astubble/wep/wep_attack.pdf
Several independent developers then released their own implementations of the FMS attack; the most popular of
these is AirSnort, which can be downloaded at the following URL:
http://airsnort.sourceforge.net/
Although traditional WLAN security that relies on open or shared keys and static WEP keys is better than no security
at all, it is not sufficient for the enterprise organization. Only very small businesses, or those that do not entrust
mission-critical data to their WLAN networks, can rely on these WLAN security types. All other enterprises and
organizations must invest in a robust, enterprise-class WLAN security solution.
Cisco agrees with the findings of the research papers discussed in the previous axiom and recommends deploying
elements of the three technologies discussed in this axiom as an alternative to WEP as specified by IEEE 802.11. The
technologies discussed include a network layer encryption approach based on IP security (IPsec), a mutual
authentication-based, key distribution method using 802.1X, and some proprietary improvements to WEP recently
implemented by Cisco. Additionally, IEEE 802.11 Task Group “i” and the Wi-Fi Alliance compliance testing
committee are working on standardizing WLAN authentication and encryption improvements.
IPsec
IPsec is a framework of open standards for ensuring secure private communications over IP networks. IPsec VPNs
use the services defined within IPsec to ensure confidentiality, integrity, and authenticity of data communications
across public networks, such as the Internet. IPsec also has a practical application to secure WLANs by overlaying
IPsec on top of cleartext 802.11 wireless traffic.
When deploying IPsec in a WLAN environment, an IPsec client is placed on every PC connected to the wireless
network and the user is required to establish an IPsec tunnel to route any traffic to the wired network. Filters are put
in place to prevent any wireless traffic from reaching any destination other than the VPN gateway and Dynamic Host
Configuration Protocol (DHCP) or Domain Name System (DNS) server. IPsec provides for confidentiality of IP
traffic, as well as authentication and anti-replay capabilities. Confidentiality is achieved through encryption using a
variant of the Data Encryption Standard (DES), called Triple DES (3DES), or the new Advanced Encryption
Standard (AES).
Though IPsec is used primarily for data confidentiality and device authentication, extensions to the standard allow
for user authentication and authorization to occur as part of the IPsec process. For more information on IPsec, refer
to the VPN primer in the SAFE VPN paper at the following URL:
http://www.cisco.com/go/safe
802.1X/EAP
An alternative WLAN security approach focuses on developing a framework for providing centralized authentication
and dynamic key distribution. This approach is based on the IEEE 802.11 Task Group “i” end-to-end framework
using 802.1X and the Extensible Authentication Protocol (EAP) to provide this enhanced functionality. Cisco has
incorporated 802.1X and EAP into its WLAN security solution—the Cisco Wireless Security Suite. The three main
elements of an 802.1X and EAP approach follow:
• Mutual authentication between client and authentication (Remote Access Dial-In User Service [RADIUS]) server
• Encryption keys dynamically derived after authentication
• Centralized policy control, where session time-out triggers reauthentication and new encryption key generation
When these features are implemented, a wireless client that associates with an access point cannot gain access to the
network until the user performs a network logon. After association, the client and the network (access point or
RADIUS server) exchange EAP messages to perform mutual authentication, with the client verifying the RADIUS
server credentials, and vice versa. An EAP supplicant is used on the client machine to obtain the user credentials (user
ID and password, user ID and one-time password [OTP], or digital certificate). Upon successful client and server
mutual authentication, the RADIUS server and client then derive a client-specific WEP key to be used by the client for
the current logon session. User passwords and session keys are never transmitted in the clear, over the wireless link.
RADIUS Server
Campus Network
Campus Network
In the Cisco SAFE wireless architecture, LEAP, EAP-TLS, and PEAP were tested and documented as viable mutual
authentication EAP protocols for WLAN deployments.
Cisco LEAP
Cisco LEAP is the widely deployed EAP type in use today in WLANs. LEAP supports all three of the 802.1X and
EAP elements mentioned previously. With LEAP, mutual authentication relies on a shared secret, the user’s logon
password, which is known by the client and the network. As shown in Figure 2, the RADIUS server sends an
authentication challenge to the client. The client uses a one-way hash of the user-supplied password to fashion a
response to the challenge and sends that response to the RADIUS server. Using information from its user database,
the RADIUS server creates its own response and compares that to the response from the client. When the RADIUS
server authenticates the client, the process repeats in reverse, enabling the client to authenticate the RADIUS server.
When this is complete, an EAP-Success message is sent to the client and both the client and the RADIUS server derive
the dynamic WEP key.
RADIUS Server
Campus Network
Campus Network
EAP-TLS
EAP-TLS is an Internet Engineering Task Force (IETF) standard (RFC 2716) that is based on the TLS protocol (RFC
2246). EAP-TLS uses digital certificates for both user and server authentication and supports the three key elements
of 802.1X/EAP mentioned previously. As shown in Figure 3, the RADIUS server sends its certificate to the client in
phase 1 of the authentication sequence (server-side TLS). The client validates the RADIUS server certificate by
verifying the issuer of the certificate—a certificate authority server entity—and the contents of the digital certificate.
When this is complete, the client sends its certificate to the RADIUS server in phase 2 of the authentication sequence
(client-side TLS). The RADIUS server validates the client’s certificate by verifying the issuer of the certificate
(certificate authority server entity) and the contents of the digital certificate. When this is complete, an EAP-Success
message is sent to the client and both the client and the RADIUS server derive the dynamic WEP key.
RADIUS Server
Campus Network
Campus Network
PEAP
PEAP is an IETF draft RFC authored by Cisco Systems, Microsoft, and RSA Security. PEAP uses a digital certificate
for server authentication. For user authentication, PEAP supports various EAP-encapsulated methods within a
protected TLS tunnel. PEAP supports the three main elements of 802.1X/EAP, as mentioned previously. As shown in
Figure 4, phase 1 of the authentication sequence is the same as that for EAP-TLS (server-side TLS). At the end of
phase 1, an encrypted TLS tunnel is created between the user and the RADIUS server for transporting EAP
authentication messages. In phase 2, the RADIUS server authenticates the client through the encrypted TLS tunnel
via another EAP type. As an example, a user can be authenticated using an OTP using the EAP-GTC subtype (as
defined by the PEAP DRAFT). In this case, the RADIUS server will relay the OTP credentials (user ID and OTP) to
an OTP server to validate the user login. When this is complete, an EAP-Success message is sent to the client and both
the client and the RADIUS server derive the dynamic WEP key. For more information on PEAP, refer to the IETF
Web site for the latest draft.
RADIUS Server
Campus Network
Campus Network
For more information on EAP protocols, implementation details, and guidelines for deployment over WLANs,
refer to:
http://www.cisco.com/warp/public/779/smbiz/wireless/wlan_security.shtml/
WEP Enhancements
Enhancements are needed to mitigate the WEP vulnerabilities discussed in the “802.11 Is Insecure” axiom section.
IEEE 802.11i includes two encryption enhancements in its draft standard for 802.11 security:
1. Temporal Key Integrity Protocol, or TKIP, which is a set of software enhancements to RC4-based WEP
2. AES, which is a stronger alternative to RC4
In December 2001, Cisco introduced support for TKIP as a component of the Cisco Wireless Security Suite. Because
the standard for TKIP was not finalized at that time, the implementation is prestandard and is sometimes referred to
as Cisco TKIP. In 2002, 802.11i finalized the specification for TKIP, and the Wi-Fi Alliance announced that it was
making TKIP a component of Wi-Fi Protected Access (WPA), which will become a requirement for Wi-Fi compliance
before the end of 2003. The enterprise version of WPA also requires 802.1X for 802.11. Both Cisco TKIP and the
http://www.weca.net
Because the most popular attack against WEP relies on exploiting multiple weak initialization vectors in a stream of
encrypted traffic using the same key, using different keys per packet is a potential way to mitigate the threat. As
illustrated in Figure 5, the initialization vector and WEP key are hashed to produce a unique packet key (called a
temporal key), which is then combined with the initialization vector and run through a mathematical function called
XOR with the plaintext. The standard 802.11 method of doing the RC4 cryptography in WEP is described in the
primer (Appendix B).
Figure 5
Per-Packet WEP Key Hashing
Ciphertext Data
HASH XOR
IV Packet Key
Stream Cipher
RC4
This scenario prevents the weak initialization vectors from being used to derive the base WEP key because the weak
initialization vectors allow you to derive only the per-packet WEP key. In order to prevent attacks due to
initialization-vector collisions, the base key should be changed before the initialization vectors repeat. Because
initialization vectors on a busy network can repeat in a matter of hours, mechanisms like EAP authentication
protocols should be used to perform the rekey operation.
Similar to a unicast key, the WLAN broadcast key (used by access points and clients for Layer 2 broadcast and
multicast communication) is susceptible to attacks due to initialization vector collisions. Cisco’s access points support
broadcast key rotation to mitigate this vulnerability. The access point dynamically calculates the broadcast WEP key
(as a function of a random number) and the new broadcast WEP key is delivered to clients using EAPOL-Key
messages. Thus, broadcast WEP key rotation can be enabled only with EAP protocols such as LEAP, EAP-TLS, and
PEAP that support dynamic derivation of encryption keys.
Another concern with WEP is its vulnerability to replay attacks. The MIC protects WEP frames from tampering. The
MIC is based on a seed value, destination MAC, source MAC, and payload (that is, any changes to these will affect
the MIC value). And the MIC is included in the WEP-encrypted payload. MIC uses a hashing algorithm to derive the
resulting value. This is an improvement of the cyclic redundancy check (CRC)-32 checksum function as performed
by standards-based WEP. With CRC-32, it is “possible to compute the bit difference of two CRCs based on the bit
Summary
Organizations should choose to deploy either IPsec or 802.1X/EAP with TKIP or Cisco TKIP, but generally not both.
Specific designs using both at the same time were tested in the SAFE lab and are discussed in the “Alternatives”
sections of the following designs. Organizations should use IPsec when they have the utmost concern for the
sensitivity of the transported data, but remember that this solution is more complex to deploy and manage than
802.1X/EAP with TKIP. The 802.1X/EAP with TKIP should be used when an organization wants reasonable
assurance of confidentiality and a transparent user security experience. The basic WEP enhancements can be used
anywhere WEP is implemented. For the vast majority of networks, the security provided by 802.1X/EAP with TKIP
is sufficient. Table 1 gives a detailed view of the pros and cons of IPsec and EAP authentication protocols in
WLAN designs:
Key length (in bits) 128 128 128 168/128, 192, 256
Client OS support Wide range Wide range Wide range Wide range
Network designers concerned about designing and implementing highly available wireless networks need to consider
both the wired and wireless elements in their design. In Cisco SAFE wireless, this paper discusses only the availability
requirements of the network elements that provide security-related services. Specifically, high availability is required
for the following three services:
• DHCP
• RADIUS
• IPsec
In wired networks, it is often possible to segment users by community through the use of Layer 3 segmentation. In
SAFE enterprise design, for example, there is a separation between a marketing segment and an R&D segment. This
segmentation occurs at the building distribution module, which is the first point of Layer 3 in the network for the
user community. Throughout the rest of the SAFE enterprise design, this segmentation can be maintained by filtering
on the IP address that the different user communities access. Furthermore, this sort of segmentation can be
administratively complex because the functional and physical separations are often two different things. For
example, a financial controller with the need to access an organization’s accounting systems could be sitting next to
a guest cubicle that needs access only to basic services.
Similar to the wired world, user differentiation in the wireless world can be implemented using wireless virtual LANs
(VLANs) mapped to wired VLANs. Mixed deployment of security standards (802.1X/EAP and IPsec VPN) are
supported through multiple VLANs, each VLAN supporting a specific security scheme. A unique wireless VLAN is
identified using a unique SSID and is mapped to a wired VLAN ID. Continuing the example, the financial controller
and guest use different SSIDs to obtain network access. Each SSID is mapped to a unique VLAN ID. Furthermore,
VLAN access control mechanisms can be implemented using the RADIUS server. For example, the access point can
dynamically map the financial controller to a VLAN ID returned by the RADIUS server upon successful 802.1X/EAP
authentication. Also, the introduction of VLANs on the access point allows organizations to separate their access
point management traffic from the normal flow of user traffic. Refer to Appendix A for an example implementation
of user differentiation (Engineering and R&D user groups) in the enterprise EAP design. For more information on
VLAN implementation on Cisco wireless platforms, refer to the “Wireless VLAN Deployment Guide” posted on
Cisco.com:
http://www.cisco.com/en/US/products/hw/wireless/ps430/prod_technical_reference09186a00801444a1.html
Without the implementation of wireless VLANs, IPsec-based VPN implementation can be used to enforce user-group
privileges. By requiring users to run a VPN client on their end hosts, you can use the wireless network purely for
transit and allow the VPN to handle any security controls. This design is discussed in detail later in the document.
Cisco SAFE wireless addresses the general concerns of WLAN security as outlined in the axiom section. This design
section integrates the concerns and mitigation techniques of the axiom section and applies them to a variety of
different networks. The size and security concerns of the specific design dictate the mitigation techniques that are
applied to a WLAN design. Therefore, the network designer is offered a choice of the mitigation technology to
implement along with the advantages and disadvantages of the technologies specific to the SAFE design. The
mitigation technologies are consistent across all the SAFE designs, so a review of the networking elements of each of
the two main technology choices is presented first. After reviewing the technologies, the network designer is presented
with each SAFE design, along with the advantages and disadvantages of implementing the specific mitigation
technologies within SAFE. Any unique characteristics of implementing a mitigation technology within the SAFE
designs are also presented. The two main design choices follow:
• Implementing a dynamic WEP keying model using 802.1X/EAP and TKIP
• Implementing an overlay VPN network using IPsec
This section outlines the generic elements of WLAN designs because so many of them are common throughout the
SAFE designs. After reading this section, you can move to the WLAN design that most interests you. In this way, the
basic concepts can be included once, with specific variances and alternatives discussed in the specific SAFE design. In
the standard WLAN designs, it is assumed that all WLAN devices are connected to a unique IP subnet to enable
end-user mobility throughout various designs. An assumption is made in the designs that most services available to
the wired network are also available to the wireless network addition. All designs include the following WLAN
security principles:
• Access point security recommendations:
– Enable centralized user authentication (RADIUS, TACACS+) for the management interface.
– Choose strong community strings for Simple Network Management Protocol (SNMP) and change them often.
– Consider using SNMP Read Only if your management infrastructure allows it.
– Disable any insecure and nonessential management protocol provided by the manufacturer.
– Utilize secure management protocols, such as Secure Shell Protocol (SSH).
– Limit management traffic to a dedicated wired subnet.
– Isolate management traffic from user traffic and encrypt all management traffic where possible.
– Enable wireless frame encryption where available.
– Physically secure the access point.
• Client security recommendations:
– Disable ad hoc mode.
– Enable wireless frame encryption where available.
This design details a generic method for using EAP with TKIP as a security mechanism to access the production
corporate network (refer to Figure 6).
Figure 6
Attack Mitigation Roles for Standard EAP WLAN Design
Threats Mitigated
• Wireless packet sniffers—Wireless packet sniffers can take advantage of any of the known WEP attacks to derive
the encryption key. These threats are mitigated by WEP enhancements (specifically per-packet keying as part of
TKIP) (see the section “Security Extensions Are Required”), and key rotation using EAP.
• Unauthenticated access—Only authenticated users are able to access the wireless and wired network. Optional
access control on the Layer 3 switch limits wired network access.
• MITM—The mutual authentication nature of several EAP authentication types combined with the MIC can
prevent hackers from inserting themselves in the path of wireless communications.
EAP-PEAP (Client
Authentication using
Attack Cisco LEAP with TKIP EAP-TLS with TKIP OTP) with TKIP
1. Use of a strong password policy is recommended to mitigate brute-force dictionary attacks against LEAP. The IT administrator also needs to limit the number
of login attempts before locking out the account.
In most cases, WLAN access points are connected to existing Layer 2 access switches. RADIUS and DHCP servers
are located in the network services module of the corporate network. Security in the design is maintained by
preventing network access to unauthenticated clients, including events of a RADIUS service failure. Because most of
the mitigation against security risks relies on the RADIUS service, this behavior is required. Overall, management of
the solution is hindered if DHCP services fail. The wireless clients and access points use EAP to authenticate the
WLAN client devices and end users against the RADIUS servers. For scalability and manageability purposes, the
WLAN client devices are configured to use the DHCP protocol for IP configuration. DHCP occurs after the device
and end user are successfully authenticated via an EAP protocol. After successful DHCP configuration, the wireless
In order to prevent attacks due to initialization vector collisions, rekeying for both unicast and broadcast keys are
recommended. For EAP with Cisco TKIP, rekeying time of 4 hours and 40 minutes is recommended for both unicast
and broadcast WEP keys. For more information, refer to white papers posted at:
http://www.cisco.com/warp/public/779/smbiz/wireless/wlan_security.shtml/
Additionally, the network designer should consider RADIUS server scalability in a large design. For example, server
load balancing products can be used to load balance between multiple RADIUS servers in a large design.
Optionally, network designers could consider implementation of unique wireless VLANs with the EAP design.
Dynamic VLAN assignment can be implemented for EAP users using the RADIUS server and user-group settings.
This has the advantage of segregating wireless users into user communities and enforcing policies for these groups at
the distribution layer. With the use of VLANs on access points, the management traffic can also be isolated from user
traffic with the implementation of management VLAN on the access points.
This design details a generic method for using IPsec VPNs as an overlay security mechanism to access the production
corporate network from a WLAN (refer to Figure 7).
WLAN access points connect to Layer 2 switches in the building module on a dedicated wired VLAN and forward
IPsec traffic from the WLAN client. The traffic is kept separate from normal wired traffic until it is decrypted by the
VPN termination device. It is important to point out that WEP is not enabled in this design. The wireless network
itself is considered an untrusted network, suitable only as a transit network for IPsec traffic. In order to isolate this
untrusted network, administrators should not mix the VLAN for the WLAN users with a wired network. This
configuration would allow hackers on the wireless network to potentially attack users on the wired network. The
WLAN clients associate with a wireless access point to establish connectivity to the campus network at Layer 2. The
wireless clients then use DHCP and DNS services in the server module to establish connectivity to the campus at
Layer 3. After the initial Layer 3 configuration, the VPN tunnel authenticates to the VPN gateway. The VPN gateway
can use digital certificates or preshared keys for wireless device authentication. If the VPN gateway uses preshared
keys for authentication, then OTPs are recommended to authenticate users to it. Without OTP, the VPN gateways
are open to brute-force login attempts by hackers who have obtained the shared IPsec key used by the VPN gateway.
The VPN gateway takes advantage of RADIUS services, which in turn contact the OTP server for user authentication.
The VPN gateway uses DHCP for IP address configuration in order for the WLAN client to communicate through
the VPN tunnel. Security in the design is maintained by preventing network access if a VPN gateway or RADIUS
service fails. Both services are required in order for the client to reach the wired network with production traffic. It
First, the access point should be configured with EtherType, protocol, and port filters based on a company's wireless
usage policy. SAFE WLAN recommends restrictive filters that allow only the necessary protocols required for
establishing a secure tunnel to a VPN gateway. These protocols include DHCP for initial client configuration; DNS
for name resolution of the VPN gateways; the VPN-specific protocols, IKE (User Datagram Protocol [UDP] port 500)
and ESP (IP Protocol 50), and ICMP for troubleshooting purposes. Even with this filtering, the DNS and DHCP
servers are still open to direct attack on the application protocols themselves. Extra care should be taken to ensure
that these systems are as secure as possible at the host level. This includes keeping them up-to-date with the latest OS
and application patches and running a host-based intrusion detection system (HIDS). Additionally, recent models of
access layer switches have the capability to implement a technology called VLAN ACL (VACL). Implementing
VACLs for VPN-related protocols and specific IP addresses of the VPN concentrators can provide an additional layer
of filtering to guarantee that only IPsec traffic destined for the appropriate enterprise VPN concentrators crosses the
switch. The DNS traffic is optional, dependent on whether the VPN client needs to be configured with a DNS
name for the VPN gateway or if only an IP address is suitable. It is recommended that ICMP be allowed only to
the outside interface of the VPN concentrator for troubleshooting purposes and path maximum-transmission-unit
(MTU) discovery.
Secondly, a VPN client feature automatically establishes a tunnel when the correct WLAN IP address is received from
DHCP. This feature eliminates the need for the end user to manually establish the VPN tunnel after the computer
startup. Third, personal firewall software is included on the wireless client to protect the client while it is connected
to the untrusted WLAN network without the protection of IPsec. In general terms, the VPN gateway delineates
between the trusted wired network and the untrusted WLAN. The wireless client establishes a VPN connection to
the VPN gateway to start secure communication to the corporate network. In the process of doing so, the VPN
gateway provides device and user authentication via the IPsec VPN. Split tunneling by the client should be
disabled—all traffic must traverse the tunnel.
Alternatives
Network designers may still consider enabling static WEP keys on all devices in an effort to add an additional
deterrent against hackers. The management overhead of dealing with static key changes makes this alternative less
than ideal for large WLAN deployments. This management overhead could be mitigated by never changing the static
WEP key, but this solution falls strongly into the “security-through-obscurity” category.
Additionally, network designers can consider using a layering of 802.1X/EAP with the IPsec-based VPN deployment
to secure their WLAN environment. The primary drawback with this alternative is the necessity of managing two
separate security infrastructures for WLAN deployments.
To further secure the DNS and DHCP services, network designers should consider using dedicated hosts for the VPN
WLAN DHCP and DNS deployment. This mitigates against two potential threats that could affect wired resources:
• DoS attacks against the DHCP and DNS services that could affect wired users
• Network reconnaissance through the use of DNS queries or reverse lookups
The large-enterprise WLAN design overlays WLANs on top of the campus portion of the SAFE enterprise blueprint.
All the components for implementing the mitigation techniques are contained within the large-enterprise building,
distribution, and server modules. These components are intended to allow WLAN access for enterprise end users
within the enterprise campus. Specifics for implementing each mitigation technique are discussed in detail in the
following sections.
Design Guidelines
In the large-enterprise WLAN design, scalability and high availability were primary concerns when implementing the
mitigation technologies. Both LEAP and VPN are considered viable security options for large-enterprise WLAN
designs. Network designers should weigh the business benefits of both technologies with the company security policy
before selecting the technology that is best suited for their network.
Network Management
In order to isolate management traffic from user traffic, it is recommended to use VLANs on the access points,
creating a management VLAN for the access points and restricting access to the access points via the management
subnet by implementing ACLs in the building distribution Layer 3 switch. The ACLs should be specific to the IP
address and protocols that the centralized multidevice access point configuration tool requires. Note that because
access points support only one wired interface, all management was done in band, versus the out-of-band
management as recommended by SAFE enterprise. This setup contains a security risk because some management
traffic (SNMP, Trivial File Transfer Protocol [TFTP], Hypertext Transfer Protocol [HTTP]) must be sent in the clear
in order to manage each access point via a central management station. The access point should be configured to
provide central authentication, authorization, and accounting (AAA) of access point administrators via RADIUS or
TACACS+, depending on the support of the deployed access points. Finally, network administrators should utilize a
secure management transport such as SSH in order to manage the access points from the command line.
Figure 8
Large-Enterprise EAP WLAN Design
Building
Module
Edge To eCommerce
Distribution Module
Building
Module
Distribution
Module
To Corporate
Internet Module
Core
Module
To VPN/Remote
Access Module
Server
Module
To WAN
Module
DHCP/AP
RADIUS/OTP/PKI
Management
Servers
Servers
EAP access via the wireless network takes advantage of three components from the SAFE enterprise architecture:
• Building module
• Building distribution module
• Server module
In the large WLAN design, the wireless access points are connected to existing Layer 2 access switches in the building
module throughout the corporate campus. RADIUS OTP, PKI, and DHCP servers are located in the server module.
The primary concern for EAP in a large WLAN design is the availability and scalability of the network configuration
and authentication servers. Following the notes in the axiom section, the RADIUS, OTP, PKI, and DHCP servers are
deployed in a redundant fashion on differing network subnets to ensure high availability and scalability. Additionally,
server load balancing products can increase the scalability of the RADIUS servers by spreading the RADIUS
authentication requests evenly across a farm of RADIUS servers. Beyond the notes listed previously, the connectivity
method is identical to the standard EAP WLAN design discussed earlier in the document.
Figure 9
Large-Enterprise VPN WLAN Design
Wireless Wireless
Computer with Computer with
VPN Client VPN Client
VPN
Concentrator
Cluster
Building
Module Edge
To eCommerce
Distribution
Module
Building Module
Distribution
Module To Corporate
Internet Module
Core
Module
To VPN/Remote
Access Module
Server
Module
To WAN
Module
DHCP/AP
RADIUS/OTP
Management
Servers
Servers
Design Guidelines
The primary objective in the large WLAN design involves balancing mitigating security risks with creating a scalable
design that a business can afford to implement. The standard VPN WLAN design guidelines in this document
outlined the general way the VPN can be implemented to secure a WLAN environment. In the context of a large
WLAN environment, the guidelines described would be cost-prohibitive for most businesses because of the
requirement for a separate Layer 2 switching infrastructure and cabling. Therefore, security trade-offs are made in
order to make a VPN WLAN feasible in a large environment. These trade-offs are noted in the following paragraphs
to help network designers decide if VPNs are a proper solution for their environment.
The WLAN clients associate with a wireless access point in the building module to establish connectivity to the
campus network at Layer 2. The wireless clients then use DHCP and DNS services in the server module to establish
connectivity to the campus at Layer 3. It should be noted that when the wireless client is communicating with the
WLAN network, but before the IPsec tunnel is established, the client traffic is not considered secure. All the noted
WLAN security issues are still present until the wireless client can secure communications with an IPsec VPN. The
auto-initiate feature of the VPN client should be utilized in order to minimize the amount of time before the VPN
tunnel is established. In addition to the filters on the access point noted in the general VPN WLAN design, the
building distribution module Layer 3 switches are configured with ACLs to permit only protocols necessary for VPN
connectivity and management. The wireless client establishes a VPN connection to the VPN gateways connecting the
building distribution and edge distribution modules. The redundant VPN gateways are configured in a
load-balancing configuration to provide high availability and scalability. These VPN gateways are a centralized
resource shared by potentially multiple Layer 2 building modules. The RADIUS, OTP, and DHCP servers used by
the VPN gateways are deployed in a redundant fashion on different network subnets within the server module to
ensure high availability and scalability of their respective services to the VPN clients tunnels.
Alternatives
An organization can further its security posture by deploying a network-based IDS (NIDS) and firewalling behind
the VPN gateways before wireless user traffic hits the production wired network. This setup allows the network to
audit, inspect, and filter user traffic that is being sent from the wireless clients to the enterprise network as defined
by an organization’s security policy. After providing the device and user authentication, the VPN gateway can
optionally provide user authorization rights based on the group with which the wireless user is associated. All these
security improvements are strongly recommended if the VPN user authentication policy chooses not to use OTP.
Also, a network designer looking for more security than the discussed design provides should consider the benefits
of building a physically separate infrastructure for WLAN access. Physically separate Layer 2 and 3 segments on
dedicated networking hardware are used to totally isolate the untrusted WLAN until traffic is decrypted at the VPN
gateways and routed into the production wired network.
The medium network WLAN overlays wireless on top of the campus portion of the SAFE medium network design.
All the components for implementing the mitigation techniques are contained within the medium-campus module.
These components are intended to allow WLAN access for end users within the medium-network campus. Specifics
for implementing each mitigation technique are discussed in detail in the following section.
Design Guidelines
In the medium WLAN design, it is assumed that all WLAN devices are connected to a single IP subnet to enable
end-user mobility throughout the medium WLAN design. An assumption is made in the designs that most services
available to the medium wired network are also available to the medium WLAN design. Keeping with the design
foundation for the SAFE medium network, the medium WLAN design does not offer high availability. Both EAP and
VPN are considered viable security options for a medium WLAN design. Key devices for both the EAP and VPN
options are supported in the campus module of the SAFE medium network design. For both options, network
designers should give special consideration to the location of the RADIUS and DHCP servers used by the EAP and
VPN WLAN solutions. The location of the servers will depend on the type of office the medium-network WLAN
represents, medium business or branch office. If the medium network is the main business office, the DHCP and
RADIUS servers will be located on the local network. If the medium network is a branch office, the DHCP and
RADIUS servers might reside at the corporate office, with connectivity via the WAN module or through a VPN in
the corporate Internet module. If the DHCP and RADIUS servers are located at the corporate office, wireless users
will be denied access to the local network if the access point or VPN gateways cannot communicate with the RADIUS
server for any reason, such as loss of WAN connectivity. Also, if the DHCP servers are unavailable to the medium
network, the wireless clients will not be able to establish IP connectivity with the campus network. Security in the
design is maintained by preventing network access if the RADIUS service fails. Because most of the mitigation against
security risks relies on the RADIUS service, this behavior is required. Overall, management of the solution is hindered
if DHCP services fail. Specifics for accomplishing these goals are detailed within each mitigation techniques section.
Network Management
In order to isolate management traffic from user traffic, it is recommended to use VLANs on the access points.
Creating a management VLAN on the access point and restricting access to the access point via the management
subnet by implementing ACLs in the building distribution Layer 3 switch is recommended. The ACLs should be
specific to the IP address and protocols that the centralized multidevice access point configuration tool require. Note
that because access points support only one wired interface, all management was done in band, versus the
out-of-band management as recommended by SAFE enterprise. This setup contains a security risk because some
Figure 10
Medium-Network EAP WLAN Design
Access Point
with EAP and TKIP
DHCP/RADIUS/OTP/PKI/AP
To WAN Management Servers
Module
EAP access in the medium WLAN design has wireless access points connected to the existing Layer 2 access switch
in the medium-campus module. RADIUS and DHCP servers are also located in the campus module, but off a distinct
Layer 3 subnet on the central-campus Layer 3 switch. The wireless EAP users will require DHCP and RADIUS
authentication services to access the medium-campus network. If the medium network is a branch office, the DHCP
and RADIUS servers may reside at the corporate office.
The process of accessing the medium network is the same as that outlined in the standard medium WLAN design
guidelines.
Alternatives
In the case of a branch office where RADIUS and DHCP servers reside at the corporate office, RADIUS and DHCP
server redundancy need to be considered. As an alternative, the network designer could choose to implement local
RADIUS and DHCP servers to provide WLAN access in the event of WAN link connectivity failure to the corporate
network. If this alternative is chosen, the network designer needs to consider administration and maintenance of
multiple (or hundreds in the case of a retail network) RADIUS and DHCP servers.
As discussed in the Large-Enterprise EAP design guidelines section, EAP design along with VLANs on access points
enables the network designer to implement user differentiation (using wired and wireless VLANs) and enforce VLAN
assignment for users and user groups using the RADIUS server. Furthermore, appropriate Layer 3 filters can be
enforced at the access and distribution layer for each user group.
Figure 11
Medium Network VPN WLAN Design
Management VPN
Servers Concentrator
To Corporate Corporate Wireless Computer
Internet Module Users with VPN Client
Access Point
DHCP/RADIUS/OTP/AP
To WAN Management Servers
Module
The IPsec VPN option in the medium network is very similar to the VPN option for the large WLAN design. The
primary differences are in the physical connectivity of the VPN gateway that divides the wireless network from the
wired network. The VPN gateway connects its interfaces to the campus-module Layer 3 switch using two different
VLANs. It should be noted that this recommendation is in direct conflict with the “Switches Are Targets” axiom in
the core SAFE security documents. When you use VLANs in a security role, you are effectively extending the security
perimeter to include the switch itself. A compromise of the switch allows the hacker to bypass the VPN concentrator.
This VLAN-based option was chosen because the alternative was not financially viable for businesses likely to deploy
a midsize network. See the following alternatives for a more secure option using additional equipment.
The VPN gateway connects its public interface to one VLAN that can connect to the wireless access points. As in the
standard VPN WLAN design, it is recommended that the VPN concentrator perform DHCP relay between the public
side and the private side of the concentrator. This allows the network designer to more effectively deploy and manage
the DHCP services for the WLAN. The private interface of the VPN gateway connects to a VLAN with access to the
wired network. The wireless access points connect to existing Layer 2 switches in the campus-module access layer
on a dedicated VLAN and forward traffic from the WLAN to the VLAN with VPN connectivity. Like the large
WLAN and general VPN WLAN design, the access and building distribution module Layer 3 switches are configured
with ACLs to permit only protocols necessary for VPN connectivity and management.
The wireless client establishes an IPsec connection to the wireless VPN gateway. In the process of doing so, the VPN
gateway provides device and user authentication via the IPsec VPN. The VPN gateway can use digital certificates or
preshared keys for wireless client device authentication. The VPN end user employs OTPs to authenticate to the VPN
gateway. The VPN gateway uses RADIUS services, which in turn contact the OTP server for user authentication. The
VPN gateway uses DHCP for IP addressing information in order for the WLAN client to communicate through the
VPN tunnel.
Figure 12
Medium-Network VPN WLAN Design
Wireless Computer
with VPN Client
Access Point
Management
Servers
To Corporate Corporate
Internet Module Users
VPN
Concentrator
DHCP/RADIUS/OTP/PKI/AP
To WAN Management Servers
Module
The small WLAN design overlays WLAN on top of the SAFE small network design. The small WLAN design is
contained within the campus module. This section discusses one option, EAP with TKIP, for providing WLAN users
connectivity to the wired campus. IPsec is not presented as an option because of the financial burden of implementing
a dedicated WLAN VPN in a network of this size.
The following sections detail the small WLAN design. Because the small-network design has a single Layer 2 switch
for its campus connectivity (as shown in Figure 13), all devices are assumed to have a single IP subnet network to
enable access point roaming.
Network Management
Network management traffic from the management hosts to the access points is unrestricted because of the lack of
a Layer 3 device in the small campus. Some management traffic is sent in the clear to each access point, as is done
for the rest of the SAFE small design.
Figure 13
Small-Network EAP WLAN Design
Corporate Corporate
Server Users
To Corporate DHCP/RADIUS/OTP/PKI
Internet Module Management Servers
Access Point
with EAP and TKIP
Wireless Computer
with EAP and TKIP
Cisco EAP access in the small WLAN design has wireless access points connected to the existing Layer 2 access switch
in the small-campus module. The wireless EAP users require DHCP and RADIUS authentication services to access
the small-campus network. Because of the single-site nature of small networks, the RADIUS and DHCP servers reside
locally connected to the Layer 2 switch in the campus module.
The process of accessing the small network is the same as that outlined in the standard WLAN design guidelines.
Alternatives
Although not recommended, if an organization is comfortable with managing the key distribution issues, static WEP
(with the cryptography fixes listed earlier) can be used as an alternative to EAP.
The remote WLAN design shows remote wireless solutions for the two primary types of remote VPN connectivity
defined by SAFE: software-based VPNs and hardware-based VPNs. This section discusses these two options for
providing WLAN users connectivity to a central office (small, medium, or enterprise) within the SAFE design.
Figure 14
Software VPN Remote Network WLAN Design
Internet
Access Point
The IPsec VPN option in the remote network is recommended when the wireless user requires security from the
wireless device to the corporate network (as illustrated in Figure 14). This is the most common configuration for
remote workers who may not have IT-managed hardware resources at their remote location. Part-time teleworkers
fall into this category. The access point can be set up with almost any configuration that allows connectivity to the
broadband device because the security is handled via the VPN client with personal firewall software. In addition,
network designers can consider putting filters on the access point to allow only IPsec, DHCP, and DNS traffic in order
to mitigate attacks from the WLAN to the wired LAN. The access point filters are detailed in Appendix A,
“Validation Lab.”
Figure 15
Hardware VPN Remote Network WLAN Design
Internet
Access Point
with EAP and TKIP
Wireless Computer
with EAP and TKIP
A reference Cisco SAFE Blueprint for network security wireless LAN (WLAN) implementation exists to validate the
functionality described in this document. This appendix details the configurations of the specific devices as they relate
to WLAN functionality within each module, as well as the overall guidelines for general device configuration. The
following are configuration snapshots from the live devices in the lab. Cisco does not recommend applying these
configurations directly to a production network.
Overall Guidelines
The sample commands presented in this section correspond in part to the SAFE WLAN design guidelines presented
earlier in this document.
Figure A-2 illustrates the security configuration for an EAP access point. “Full Encryption” (wired equivalent privacy
[WEP]) is mandated by the access point; in addition, the access point allows network EAP as the only authentication
method. EAP types configured for this deployment included Cisco-EAP (LEAP), Protected EAP (PEAP), and
EAP-Transport Layer Security (EAP-TLS).
Figure A-2
WEP Configuration for an EAP Access Point
As Figure A-3 illustrates, an access point is configured to allow open authentication, and WEP encryption is not
enabled for virtual-private-network (VPN) wireless clients authenticating to a wired network.
The following sections detail sample configurations needed on access points to enable EAP or VPN (shown in screen
captures) as detailed in the axioms and design guidelines sections of this document. The sample configuration screen
captures were taken for the large-enterprise design.
As discussed in the design section of this document, a VPN access point should be configured with EtherType,
protocol, and port filters based on a company’s wireless usage policy. SAFE WLAN recommends restrictive filters
that allow only the necessary protocols required for establishing a secure tunnel to a VPN gateway. The Internet
Control Message Protocol (ICMP) IP protocol filter is allowed for troubleshooting purposes. Tables A-1 and A-2 list
the inbound (receive) and outbound (transmit) filters to be set on the VPN access point radio interface:
The following sections detail sample configurations needed for wireless clients to enable VPN or EAP (shown in
screen captures) as detailed in axioms and design guidelines sections of this document. The sample configuration
screen captures were taken for the large-enterprise design. However, configuration for a VPN wireless client (or a
LEAP wireless client) is identical for all designs.
VPN Client
For a wireless user connecting to a wired network using a VPN client, encryption and authentication at the Layer 2
level are usually disabled. However, the IT administrator could choose to use static-WEP or Temporal Key Integrity
Protocol (TKIP) or EAP with TKIP at the Layer 2 network level. Figures A-4 and A-5 illustrate the sample setup.
Figure A-5
Network Security Configuration on a VPN Client
EAP Client
A wireless client is configured for EAP by specifying the appropriate service set ID (SSID) and security settings using
the Cisco Aironet® Client utility. However, it should be noted that OS-level configuration is usually required for
EAP-TLS and PEAP (for example, configuring a Microsoft Windows XP client with the appropriate EAP-TLS or
PEAP settings). Figures A-6, A-7, and A-8 illustrate sample configuration for a LEAP client. Figures A-9 and A-10
illustrate sample configuration for an EAP-TLS or PEAP client.
Figure A-7
Network Security Configuration for LEAP Wireless Clients
Figure A-9
Network Security Configuration for PEAP or EAP-TLS Wireless Clients
This section details end-to-end EAP and VPN architecture configurations for a large-enterprise network design.
Wireless Wireless
Computer Computer
with EAP with EAP
and TKIP and TKIP
Core
Module eL3sw-3 eL3sw-4
To VPN/Remote
eL3sw-8 Access Module
Server
Module eL3sw-1 eL3sw-2
To WAN
WLSE Module
The following sections discuss configurations specific to the SAFE WLAN large-network design. For generic
configuration guidelines for a large enterprise network, refer to “Cisco SAFE: A Security Blueprint for Enterprise
Networks.”
Refer to the configuration samples in the “Overall Guidelines” section of this appendix for configuring Cisco Aironet
access points and wireless clients for EAP. In the SAFE architecture lab, the large-enterprise EAP WLAN design was
implemented with wireless VLANs. This section details the implementation details specific to VLAN implementation
along with EAP WLAN design.
As Figures A-12 and A-13 illustrate, an access point is configured with multiple security profiles, which are correlated
with SSID associations and VLAN IDs. An EAP authenticated user is configured for full 128-bit encryption.
Figure A-12
EAP Access Point VLAN Summary Page
As discussed in the design section, marketing and engineering VLANs were created to correspond with user groups
on the Cisco Secure ACS server. Thus, when EAP wireless users associate, they are temporarily placed in VLAN 73,
and the authentication request is sent to the ACS server. When the users have been authenticated based on the
configured username, they are placed in the appropriate marketing or engineering VLAN. A management VLAN has
also been created (VLAN 70) for access point management. Refer to the VLAN deployment guide on Cisco.com for
more details:
http://www.cisco.com/en/US/products/hw/wireless/ps430/prod_technical_reference09186a00801444a1.html
In addition, an alternative configuration is to create guest (eEAP-Guest) and static WEP (eEAP-Static) SSIDs that are
associated with separate VLANs and security profiles to differentiate user access.
Figure A-13
EAP Access Point Service Set Summary Page
The following are ACLs for the management VLAN (Interface 70):
! Permit only authentication and accounting requests from APs to RADIUS server on the management
VLAN
access-list 170 permit udp host 10.1.70.120 gt 1023 host 10.1.20.54 eq 1645
access-list 170 permit udp host 10.1.70.120 gt 1023 host 10.1.20.54 eq 1646
access-list 170 permit udp host 10.1.70.121 gt 1023 host 10.1.20.54 eq 1645
access-list 170 permit udp host 10.1.70.121 gt 1023 host 10.1.20.54 eq 1646
! Permit only SNMP and TFTP traffic from wireless APs to WLSE* in out of band management network
access-list 170 permit udp host 10.1.70.120 eq snmp host 10.1.20.150
access-list 170 permit udp host 10.1.70.121 eq snmp host 10.1.20.150
access-list 170 permit udp host 10.1.70.120 host 10.1.20.150 eq tftp
access-list 170 permit udp host 10.1.70.121 host 10.1.20.150 eq tftp
! Permit outgoing traffic for AP web management in the out of band management network
access-list 170 permit tcp host 10.1.70.120 eq www 10.1.20.0 0.0.0.255 gt 1023 established
access-list 170 permit tcp host 10.1.70.121 eq www 10.1.20.0 0.0.0.255 gt 1023 established
!Permit SSH traffic from the APs to out of band management network
access-list 170 permit tcp host 10.1.70.120 eq 22 10.1.20.0 0.0.0.255 gt 1023 established
access-list 170 permit tcp host 10.1.70.121 eq 22 10.1.20.0 0.0.0.255 gt 1023 established
! Permit only SNMP and TFTP traffic from WLSE* to APs in out of band management network
access-list 171 permit udp host 10.1.20.150 gt 1023 host 10.1.70.120 eq snmp
access-list 171 permit udp host 10.1.20.150 gt 1023 host 10.1.70.121 eq snmp
access-list 171 permit udp host 10.1.20.150 gt 1023 host 10.1.70.120 eq tftp
access-list 171 permit udp host 10.1.20.150 gt 1023 host 10.1.70.121 eq tftp
! Permit outgoing web traffic from out of band network to APs for management
access-list 171 permit tcp 10.1.20.0 0 .0.0.255 gt 1023 host 10.1.70.120 eq www
access-list 171 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.70.121 eq www
access-list 171 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.70.120 eq 22
access-list 171 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.70.121 eq 22
The following are ACLs for the EAP VLAN (Interface 73):
!Deny user access to APs management VLAN
access-list 172 deny ip 10.1.73.0 0.0.0.255 10.1.70.0 0.0.0.255 log
! Permit all IP traffic from EAP VLAN in wireless network to any destination
access-list 172 permit ip 10.1.73.0 0.0.0.255 any
! Permit only BOOTP requests to pass through to DHCP server
access-list 172 permit udp host 0.0.0.0 eq bootpc host 255.255.255.255 eq bootps
!Deny all other traffic
access-list 172 deny ip any any log
VPN Architecture
The following shows a configuration snapshot from the SAFE large-enterprise VPN WLAN design. Figure A-14
illustrates the VPN design for a WLAN in a large enterprise.
Figure A-14
Large-Enterprise VPN WLAN Design
Wireless Wireless
Computer with Computer with
VPN Client VPN Client
eVPN-49
eAP1200E-122 eAP350V-123 Edge
Distribution
Building eL2sw-12 Module
Module eL2sw-11
To eCommerce
Module
Building eVPN-50
Distribution eL3sw-5
eL3sw-6 To Corporate
Module eL3sw-7
Internet Module
Core
Module eL3sw-3 eL3sw-4
To VPN/Remote
eL3sw-8 Access Module
Server
Module eL3sw-1 eL3sw-2
To WAN
WLSE Module
The following sections discuss conf igurations specific to the SAFE WLAN large-enterprise VPN network design. For
generic configuration guidelines for a large-enterprise network, refer to “Cisco SAFE: A Security Blueprint for
Enterprise Networks.”
Cisco Aironet 1200 Access Points (product numbers eAP1200V-120 and eAP350V-121) and Wireless Clients:
Refer to the configuration samples in the “Overall Guidelines” section of this appendix for configuring Cisco Aironet
access points and wireless clients for VPN connectivity over a WLAN. In the SAFE architecture lab, the
large-enterprise VPN WLAN design was implemented with wireless VLANs. This section details the implementation
details specific to VLAN implementation along with the VPN-over-WLAN design.
As Figure A-15 illustrates, an access point is configured with multiple security profiles, which are correlated with
SSID associations and VLAN IDs. As stated in the “Standard VPN design guidelines” section, Layer 2 encryption
and authentication is usually disabled for IP security (IPsec) VPN-over-WLAN deployments. However, if the IT
administrator wishes, Layer 2 encryption and authentication can be optionally enabled. In the SAFE architecture lab,
the large VPN design was implemented with Layer 2 encryption enabled. A VPN user is configured to have full
128-bit WEP encryption with Cisco TKIP and message integrity check (MIC) enabled.
Figure A-15
VPN Access Point VLAN Summary Page
As discussed in the design section, VPN users can be placed in a separate VLAN based on the SSID with which they
are associated. A management VLAN has also been created (VLAN 80) for access point management. Refer to the
VLAN deployment guide on Cisco.com for more details:
http://www.cisco.com/en/US/products/hw/wireless/ps430/prod_technical_reference09186a00801444a1.html
In addition, an alternative configuration is to create guest (eVPN-Guest) and static WEP (eVPN-Static) SSIDs that
are associated with separate VLANs and security profiles to differentiate user access (as shown in Figure A-16).
Cisco Catalyst 6506 Layer 3 Switches (product numbers eL3sw-5 and eL3sw-6) (WLAN building
module-to-building distribution module interconnection):
! Management VLAN for APs in the campus network
interface Vlan80
ip address 10.1.80.5 255.255.255.0
ip access-group 180 in
ip access-group 181 out
ip helper-address 10.1.11.50
no cdp enable
!The following are ACLs for the management VLAN (Interface 80):
! Permit only authentication and accounting requests from AP to RADIUS server
access-list 180 permit udp host 10.1.80.122 gt 1023 host 10.1.20.54 eq 1645
access-list 180 permit udp host 10.1.80.122 gt 1023 host 10.1.20.54 eq 1646
access-list 180 permit udp host 10.1.80.123 gt 1023 host 10.1.20.54 eq 1645
access-list 180 permit udp host 10.1.80.123 gt 1023 host 10.1.20.54 eq 1646
! Permit only SNMP and TFTP traffic from wireless APs to WLSE* in out of band management network
access-list 180 permit udp host 10.1.80.122 eq snmp host 10.1.20.150
access-list 180 permit udp host 10.1.80.123 eq snmp host 10.1.20.150
access-list 180 permit udp host 10.1.80.122 host 10.1.20.150 eq tftp
access-list 180 permit udp host 10.1.80.123 host 10.1.20.150 eq tftp
!Permit SSH traffic from the APs to out of band management network
access-list 180 permit tcp host 10.1.80.122 eq 22 10.1.20.0 0.0.0.255 gt 1023 established
access-list 180 permit tcp host 10.1.80.123 eq 22 10.1.20.0 0.0.0.255 gt 1023 established
! Permit incoming web traffic from out of band management network to APs
access-list 181 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.80.122 eq www
access-list 181 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.80.123 eq www
access-list 181 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.80.122 eq 22
access-list 181 permit tcp 10.1.20.0 0.0.0.255 gt 1023 host 10.1.80.123 eq 22
! Permit only SNMP and TFTP traffic from WLSE to wireless APs
access-list 181 permit udp host 10.1.20.150 gt 1023 host 10.1.80.122 eq snmp
access-list 181 permit udp host 10.1.20.150 gt 1023 host 10.1.80.123 eq snmp
access-list 181 permit udp host 10.1.20.150 gt 1023 host 10.1.80.122 eq tftp
access-list 181 permit udp host 10.1.20.150 gt 1023 host 10.1.80.123 eq tftp
! Deny all other traffic, don't log Windows file share broadcasts
access-list 182 deny udp 10.1.83.0 0.0.0.255 any eq netbios-ns
access-list 182 deny udp 10.1.83.0 0.0.0.255 any eq netbios-dgm
access-list 182 deny ip any any log
! Permit DHCP responses for the initial IP assignment for the wireless client
access-list 183 permit udp host 10.1.11.50 eq bootps host 255.255.255.255 eq bootpc
access-list 183 permit udp host 10.1.11.50 eq bootps 10.1.83.0 0.0.0.255 eq bootpc
*Cisco Wireless LAN Solution Engine (WLSE) is a wireless network management appliance that enables
configuration and management of a wireless LAN of up to 500 access points per device. Cisco WLSE was
incorporated in the SAFE design for ease of configuration and management of all access points.
Please note that the Cisco Discovery Protocol was enabled because it is part of the out-of-band management network.
This section details end-to-end EAP and VPN architecture configurations for a medium-enterprise network.
Figure A-17
Medium-Enterprise EAP WLAN Design
mCAT-6 mAP1100E-122
mCAT-5
To WAN
mIDS-3 DHCP/ACS/CA Servers
Module
The following sections discuss configurations specific to SAFE WLAN medium-network design (with EAP option).
For generic configuration guidelines for a medium-enterprise network, refer to “SAFE: Extending the Security
Blueprint to Small, Midsize, and Remote-User Networks.”
Cisco Aironet Access Point and Client (product number mAP1100E-122) and Wireless Client:
Refer to the configuration samples in the “Overall Guidelines” section of this appendix for configuring access points
and wireless clients for the EAP option. In the SAFE architecture lab, the medium-enterprise EAP WLAN design was
implemented with wireless VLANs. This section details the implementation details specific to VLAN implementation
along with the EAP WLAN design.
As shown in the output, the access points in the medium design were also configured with multiple SSIDs (product
numbers mEAP-Guest, mEAP-Static, etc.) that were associated with specific VLANs and security profiles for user
differentiation.
! Permit outgoing web and SSH management traffic from AP to the out of band management network
access-list 170 permit tcp host 10.3.70.120 eq www 10.3.8.0 0.0.0.255 gt 1023 established
access-list 170 permit tcp host 10.3.70.120 eq 22 10.3.8.0 0.0.0.255 gt 1023 established
! Permit DHCP responses for the initial IP assignment for the wireless client
access-list 173 permit udp host 10.3.2.50 eq bootps host 255.255.255.255 eq bootpc
! Permit all IP traffic from any destination to EAP VLAN in wireless network
access-list 173 permit ip any 10.3.73.0 0.0.0.255
!Deny all other traffic
access-list 173 deny ip any any log
Figure A-18
Medium-Enterprise VPN WLAN Design
Management
Servers mVPN-2
To Corporate Corporate Wireless Computer
Internet Module Users with VPN Client
mCAT-6 mAP350V-123
mCAT-5
DHCP/ACS
To WAN Servers
mIDS-3
Module
Cisco Aironet access point and client (product number mAP350V-123) and wireless client
Refer to the configuration samples in the “Overall Guidelines” section of this appendix for configuring access points
and wireless clients for the VPN option.
Cisco VPN 3000 Series concentrator (product number mVPN-2) was configured as follows: The public interface was
on the same VLAN as the VPN user VLAN (mVPN-VPN), the private interface was on a separate VLAN, and the
VPN concentrator was configured with DHCP relay enabled to forward DHCP requests from WLAN clients to the
DHCP server in the wireless network. Thus, the Catalyst Layer-3 switch (product number MCAT-6) does not require
any Layer 3 configurations. Figures A-19 and A-20 below illustrate the VPN concentrator configuration.
Figure A-20
VPN Concentrator Configuration
Configurations for the small and remote designs are not provided because there are no unique configuration elements
for these designs. Refer to the generic recommendations provided at the beginning of this section for guidance.
Standard 802.11-based wireless LANs (WLANs) provide mobility to network users while maintaining the requisite
connectivity to corporate resources. As laptops become more pervasive in the workplace, users are more prone to use
laptops as their primary computing device, allowing greater portability in meetings and conferences and during
business travel. WLANs offer organizations greater productivity per employee by providing constant connectivity to
traditional networks in venues where previously unavailable.
Wireless network connectivity is not limited to enterprise use. WLANs offer increased productivity not only before
and after meetings, but also outside the traditional office environment. Numerous wireless Internet service providers
(WISPs) are appearing in airports, coffee shops, hotels, and conference and convention centers, enabling enterprise
users to connect in public access venues.
Wireless local-area networking has existed for many years, providing connectivity to wired infrastructures where
mobility was a requirement to specific working environments. These early networks were based on both
frequency-hopping and direct-sequencing radio technologies (described later). These early wireless networks were
nonstandard implementations, with speeds ranging between 1 and 2 MB. Without any standards driving WLAN
technologies, the early implementations of WLAN were relegated to vendor-specific implementation, with no
provision for interoperability, inhibiting the growth of standards-based WLAN technologies. Today, several
standards exist for WLAN applications: 802.11, HiperLAN, HomeRF SWAP, and Bluetooth.
Functional View
From a functional viewpoint, WLANs can be categorized as follows: peer-to-peer wireless LANs, multiple-cell
wireless LANs, and building-to-building wireless networks (point to point and point to multipoint). In a peer-to-peer
wireless LAN, wireless clients equipped with wireless network interface cards (NICs) communicate with each other
without the use of an access point. Coverage area is limited in a peer-to-peer LAN, and wireless clients do not have
access to wired resources. A multiple-cell wireless LAN extends the coverage through the use of overlapping cells.
Coverage area of a cell is determined by the characteristics of the access point (a wireless bridge) that coordinates the
wireless clients’ use of wired resources.
Building-to-building wireless networks address the connectivity requirement between LANs (buildings) in a
campus-area network. There are two different types of building-to-building wireless networks: point to point and
point to multipoint. Point-to-point wireless links between buildings are radio- or laser-based point-to-point links. A
radio-based point-to-point bridged link between buildings uses directional antennas to focus the signal power in a
narrow beam, maximizing the transmission distance. A laser-based point-to-point bridged link between buildings
uses laser light (usually infrared light) as a carrier for data transmission. A radio-based point-to-multipoint bridged
network uses antennas with wide beam width to connect multiple buildings (LANs) in a campus-area network.
HiperLAN
HiperLAN is a European Telecommunications Standards Institute (ETSI) standard ratified in 1996. HiperLAN/1
standard operates in the 5-GHz radio band up to 24 Mbps. The ETSI has recently approved HiperLAN/2, which
operates in the 5-GHz band at up to 54 Mbps using a connection-oriented protocol for sharing access among
end-user devices.
HomeRF SWAP
In 1988, The HomeRF SWAP Group published the Shared Wireless Access Protocol (SWAP) standard for wireless
digital communication between PCs and consumer electronic devices within the home. SWAP supports voice and
data over a common wireless interface at 1 and 2-Mbps data rates using frequency-hopping and spread-spectrum
techniques in the 2.4-GHz band.
Bluetooth
Bluetooth is a personal-area network (PAN) specified by the Bluetooth Special Interest Group for providing
low-power and short-range wireless connectivity using frequency-hopping spread spectrum in the 2.4-GHz frequency
environment.
The IEEE maintains the 802.11-based standard, as well as other 802-based networking standards, such as 802.3
Ethernet. A nonprofit, vendor-neutral organization known as the Wi-Fi Alliance provides a branding for
802.11-based technology known as Wi-Fi. A Wi-Fi compliant device must pass interoperability testing in the Wi-Fi
laboratory. All vendor products that are Wi-Fi certified are guaranteed to work with all other Wi-Fi certified
products—regardless of the vendor.
Standard 802.11-based wireless technologies take advantage of the radio spectrum deemed usable by the public. This
spectrum is known as the Industrial, Scientific, and Medical (ISM) band. The 802.11 standard specifically takes
advantage of two of the three frequency bands, the 2.4 GHz-to-2.4835 GHz UHF band used for 802.11 and 802.11b
networks, and the 5.15 GHz-to-5.825 GHz SHF band used for 802.11a-based networks.
The spectrum is classed as unlicensed, meaning there is no one owner of the spectrum, and anyone can use it as long
as that user’s device complies with FCC regulations. Some of the areas the FCC governs include the maximum
transmit power of the radios and the type of encoding and frequency modulations that can be used.
The 2.4-GHz ISM band (used by 802.11b) makes use of spread-spectrum technology. Spread spectrum dictates that
data transmissions are spread across numerous frequencies. The reason for this is that the 2.4-GHz band has other
primary owners. Primary owners are entities who have bought the spectrum for their own use, or have been granted
legal access to the spectrum above all else. Common primary owners of the 2.4-GHz band include microwave oven
manufacturers. Microwave ovens transmit in the same frequency range, but at far greater power levels (a typical
The 802.11 standard specifies two different types of Layer 1 physical interfaces for radio-based devices. One uses a
frequency-hopping architecture, whereas the other uses a more straightforward single-frequency approach, known
as direct sequencing.
Frequency Hopping
The 2.4-GHz ISM band provides for 83.5 MHz of available frequency spectrum. The frequency-hopping architecture
makes use of the available frequency range by creating hopping patterns to transmit on one of 79 1-MHz-wide
frequencies for no more than 0.4 seconds at a time (refer to Figure B-1). This setup allows for an interference-tolerant
network. If any one channel stumbles across an interference, it would be for only a small time slice because the
frequency-hopping radio quickly hops through the band and retransmits data on another frequency.
The major drawback to frequency hopping is that the maximum data rate achievable is 2 Mbps. Although you can
place frequency-hopping access points on 79 different hop sets, mitigating the possibility for interference and
allowing greater aggregated throughput, scalability of frequency-hopping technologies becomes a deployment issue.
Work is being done on wide-band frequency hopping, but this concept is not currently standardized with the IEEE.
Wide-band frequency hopping promises data rates as high as 10 Mbps.
Figure B-1
Frequency Hopping
100 mw/MHz
Power
1 Sec
Time
1 mw/MHz 1 ms
2.402 GHz Frequency 2.483 GHz
Figure B-2
Direct Sequencing
100 mw/MHz
Power
e e
Us Us 1 Sec
n n
o ti o ti
elN elN
a nn a nn
Ch Ch Time
1 mw/MHz 1 ms
22
2.402 GHz Frequency 2.483 GHz
802.11a Networks
In 1999, the IEEE also ratified another Layer 1 physical interface, known as 802.11a. The 802.11a standard uses the
5-GHz SHF band to achieve data rates as high as 54 Mbps.
Unlike the 802.11 and 802.11b standards, the 802.11a standard uses a type of frequency-division multiplexing
(FDM) called orthogonal FDM (OFDM). In a FDM system, the available bandwidth is divided into multiple data
carriers. The data to be transmitted is then divided among these subcarriers. Because each carrier is treated
independent of the others, a frequency guard band must be placed around it. This guard band lowers the bandwidth
The FCC has broken the 5-GHz spectrum into three parts, as part of the Unlicensed National Information
Infrastructure (U-NII). Each of the three U-NII bands has 100 MHz of bandwidth and consists of four
nonoverlapping channels that are 20 MHz wide. As a result, each of the 20-MHz channels comprises 52
300 kHz-wide subchannels. Forty-eight of these subchannels are used for data transmission, while the remaining four
are used for error correction. Three U-NII bands are available for use:
• U-NII 1 devices operate in the 5.15- to 5.25-GHz frequency range. U-NII 1 devices have a maximum transmit
power of 50 mW, a maximum antenna gain of 6 dBi, and the antenna and radio are required to be one complete
unit (no removable antennas). U-NII 1 devices can be used only indoors.
• U-NII 2 devices operate in the 5.25- to 5.35-GHz frequency range. U-NII 2 devices have a maximum transmit
power of 250 mW and maximum antenna gain of 6 dBi. Unlike U-NII 1 devices, U-NII 2 devices may operate
indoors or outdoors, and can have removable antennas. The FCC allows a single device to cover both U-NII 1
and U-NII 2 spectra, but mandates that if used in this manner, the device must comply with U-NII 1 regulations.
• U-NII 3 devices operate in the 5.725- to 5.825-GHz frequency range. These devices have a maximum transmit
power of 1W and allow for removable antennas. Unlike U-NII 1 and U-NII 2 devices, U-NII 3 devices can operate
only in outdoor environments. As such, the FCC allows up to a 23-dBi gain antenna for point-to-point
installations, and a 6-dBi gain antenna for point-to-multipoint installations.
To provide some perspective on 802.11 station roaming, we first review 802.3 Ethernet network architecture.
Standard 802.3-based Ethernet LANs use the carrier sense multiple access collision detect (CSMA/CD) architecture.
A station that wishes to transmit data to another station first checks to see if the medium is in use—the carrier sense
function of CSMA/CD. All stations that are connected to the medium have equal access to it—the multiple access
portion of CSMA/CD. If a station verifies that the medium is available for use, it begins transmitting. If two stations
sense that the medium is available and begin transmitting at the same time, their frames will “collide” and render the
data transmitted on the medium useless. The sending stations are able to detect a collision, the collision detection
function of CSMA/CD, and run through a fallback algorithm to retransmit the frames.
The 802.3 Ethernet architecture was designed for wired networks. The designers placed a certain amount of
reliability on the wired medium to carry the frames from a sender station to the desired destination. For that reason,
802.3 has no mechanism to determine if a frame has reached the destination station. The 802.3 standard relies on
upper-layer protocols to deal with frame retransmission.
Standard 802.11 networks transmit across the air, and are subject to numerous sources of interference. The designers
of 802.11 understood this issue, and provided a link layer acknowledgment function to provide notifications to the
sender that the destination has received the frame. For every frame transmitted, the receiving station responds with
an acknowledgment (ACK) frame.
Access points also send out periodic management frames known as beacons. Beacons contain access point
information such as the service set identifier (SSID), support data rates, whether the access point supports frequency
hopping or direct sequencing, and capacity. Beacon frames are broadcast from the access point at regular intervals,
adjustable by the administrator.
ACK frames and beacons provide the client station with a reference point to determine whether a roaming decision
needs to be made. If a set number of beacon messages are missed, the client can assume they have roamed out of
range of the access point with which they are associated. In addition, if expected ACK messages are not received,
clients can also make the same assumption.
The actual act of roaming can differ from vendor to vendor. The basic act of roaming is making a decision to roam,
followed by the act of locating a new access point to roam to. This scenario can involve reinitiating a search for an
access point, in the same manner the client would when it is initialized, or other means, such as referencing a table
built during the previous association.
The timing of WLAN roams also varies according to vendor, but in most cases is less than 1 second, and in the best
cases, less than 200 msec. It is also important to note that because roaming is vendor specific, roaming between
different vendors’ access points can have extended roam times.
Wireless Security
As standardized by the IEEE, security for 802.11 networks can be simplified into two main components: encryption
and authentication. The implementation of these components has been proven and documented as insecure by the
security community at large. They are presented here so the reader can understand the fundamental flaws when they
are presented in the axioms section of this document.
Frame Encryption
Properly performed encryption allows for confidentiality. Encryption is the process of taking a message, referred to
as cleartext, and passing it through a mathematical algorithm to produce what is known as ciphertext. Decryption
is the reverse of the process. Encryption algorithms typically rely on a value, called a key, in order to encrypt and
decrypt the data. Two major forms of encryption are used today—symmetric encryption (also known as shared-key
encryption) and asymmetric encryption (also know as public or private encryption). Symmetric encryption is about
1000 times faster than asymmetric encryption, and is, therefore, used for the bulk encryption of data. Generally with
well-designed encryption algorithms, longer keys result in a higher degree of security because more brute force is
required to try every possible key (known as the key space) in order to decrypt a message. The IEEE has specified
that wired equivalent privacy (WEP) be the means to encrypt 802.11 data frames. WEP uses the RC4 stream cipher
that was invented by Ron Rivest of RSA Data Security, Inc. (RSADSI) for encryption. The RC4 encryption algorithm
is a symmetric stream cipher that supports a variable-length key. A stream cipher is one that operates the encrypt or
decrypt function on a unit of plaintext (in this case, the 802.11b frame). This cipher is contrasted with a block cipher,
which processes a fixed number of bytes in each encrypt or decrypt function. With symmetric encryption, the key is
the one piece of information that must be shared by both the encrypting and decrypting endpoints. RC4 allows the
Because WEP is a stream cipher, a mechanism is required to ensure that the same plaintext will not generate the same
ciphertext. The IEEE stipulated the use of an initialization vector to be concatenated with the symmetric key before
generating the stream ciphertext.
The initialization vector is a 24-bit value (ranging from 0 to 16777215). The IEEE suggests—but does not
mandate—that the initialization vector change per frame. Because the sender generates the initialization vector with
no standard scheme or schedule, it must be sent to the receiver unencrypted in the header portion of the 802.11 data
frame. The receiver can then concatenate the received initialization vector with the WEP key (base key) it has stored
locally to decrypt the data frame. As illustrated in Figure B-3, the plaintext itself is not run through the RC4 cipher,
but rather the RC4 cipher is used to generate a unique keystream for that particular 802.11 frame using the
initialization vector and base key as keying material. The resulting unique keystream is then combined with the
plaintext and run through a mathematical function called XOR. This produces the ciphertext.
Figure B-3
WEP Encryption Process
STREAM CIPER
Authentication Mechanism
The IEEE specified two authentication algorithms for 802.11-based networks. First, open authentication is a null
authentication algorithm because any station requesting authentication is granted access. The second form of
authentication is called shared-key authentication, which requires that both the requesting and granting stations be
configured with matching WEP keys. The requesting stations send an authentication request to the granting station.
The granting station sends a plaintext challenge frame to the requesting station. The requesting station WEP encrypts
the challenge frame and sends it back to the granting station. The granting station attempts to decrypt the frame, and
if the resulting plaintext matches what the granting station originally sent, then the requesting station has a valid key
and is granted access.
Note that shared-key authentication has a known flaw in its concept. Because the challenge packet is sent in the clear
to the requesting station and the requesting station replies with the encrypted challenge packet, an attacker can derive
the stream cipher by analyzing both the plaintext and the ciphertext. This information can be used to build
decryption dictionaries for that particular WEP key. The known insecurities with WEP as standardized in the IEEE
are discussed in the axioms section of this document.
As discussed in the document, the threat posed by rogue access points can be mitigated. This appendix provides
detailed guidance on minimizing the risk that rogue access points represent to enterprise networks using the
following:
Prevention
• Corporate policy
• Physical security
• Supported WLAN infrastructure
• Standard 802.1X port-based security on edge switches
Detection
• Using wireless analyzers or sniffers
• Using scripted tools on the wired infrastructure
• Physically observing WLAN access point placement and usage
At the end of the appendix is a listing of wireless sniffers and known Media Access Control (MAC) addresses. This
information can assist in the detection and prevention of rogue access points.
An enterprise policy concerning WLAN installations is an essential first step in preventing rogue access points. The
WLAN policy should include who is authorized to install WLAN access points, and what security policies must be
complied with when they are installed. For instance, the policy should dictate which security scheme the access point
must utilize for secure client connectivity.
Physical Security
Physical security also plays a part in rogue access point prevention. Physical security standards should be in place to
prevent an intruder from gaining unauthorized access to the enterprise premises or to detect the intruder if physical
access is gained.
Given that almost all rogue access points are installed by a nonmalicious (frustrated insider) class of users, the best
way to prevent such rogue installs is to remove the motivation for them. Installing a managed, supported, and secure
WLAN network throughout the enterprise removes the motivation for employees to install rogue access points.
Frustrated insiders are categorized as those who install an unauthorized access point in order to provide wireless
coverage where none is officially available; for example, enabling wireless networking in a meeting room, cafeteria,
outdoor space, or other common area. The wide availability of low-cost access points has made this installation type
very easy. The threat posed by this class of install is that the person installing the access point is often ignorant of
security features that are necessary to prevent outsiders from accessing the enterprise network, and the consumer
grade access point commonly used in this installation does not have the features to provide an enterprise level of
security.
This section looks at ways in which Layer 2 and 3 switch features might be used to prevent rogue access points. Table
C-1 lists the methods that are investigated.
Using Cisco Catalyst® switch filters to Limiting the number of MAC addresses that can be used per port can
limit MAC addresses per port prevent switches from passing traffic from rogue access point clients.
Using Cisco Catalyst switch filters to The possibility of configuring the switch to drop frames from vendors of
drop frames from third-party access third-party access points and third-party wireless network interface cards
points (NICs) was investigated.
This option is not feasible because switches do not support wild-card
masks on MAC address filters.
Limitations of Using Filters to Prevent Rogue Access Points from Connecting to the Enterprise
Network
• Filters designed to prevent a vendor’s access points will also prevent that vendor’s NICs. For example, preventing
vendor A access points could also prevent vendor A Ethernet NICs.
• Traffic coming through an access point has the MAC address of the wireless NIC, not of the access point. For
example, if an access point from vendor A is connected to a port with a vendor A MAC filter on it, it would still
be able to pass traffic from a vendor B or vendor C wireless NIC. The resolution for this issue would be to use a
command (if wildmask MACs were configurable) to shut down the port if an unauthorized MAC was seen.
• WLAN access point vendors will come and go, and there will never be an authoritative list of all access point
vendors’ MAC Organizational Unique Identifiers (OUIs).
• MAC addresses can be changed or spoofed.
In addition to the rogue access point prevention mechanisms mentioned in the previous section, a combination of the
following rogue access point detection methods should also be used by the IT security administrator:
• Detecting rogue access points wirelessly
• Detecting rogue access points from the wired network
• Detecting rogue access points by physical observation
Detecting rogue access points wirelessly is the process of using WLAN hardware and software to detect rogue
access points.
Table C-2 Advantages and Disadvantages of Wireless Detection of Rogue Access Points
• Often picks up access points that the • This method requires that the WLAN device be within range of an
other rogue access point detection access point to be able to detect it and may require that the IT
methods miss security administrator perform a walkthrough of the enterprise
• Is very effective at detecting access campus with a wireless analyzer.
points installed by the nonmalicious • Many tools do not see access points that do not broadcast their SSID.
(frustrated insider) class of installer • IT security administrator cannot easily survey remote sites.
(default security options or broadcast • WLAN access point signals may be difficult to pick up because of
service set identifier [SSID]) building materials that block 802.11 signals.
Many WLAN analyzers are available, and to various degrees they are capable of detecting rogue access points. See
section “Wireless Analyzers” in this appendix for a listing of some of the wireless analyzers that can be used for
detecting rogue access points. When a WLAN analyzer detects a suspected rogue access point, a direction antenna
on the analyzer is a very useful aid in locating the access point.
Rogue access points can be detected from the wired network using:
• MAC addresses
• OS fingerprinting
• Simple Network Management Protocol (SNMP)
• Intrusion detection
A large number of software tools are available to aid in detecting rogue access points from a wired management
station on the Ethernet portion of the network.
Table C-3 lists the advantages and disadvantages of wired detection of rogue access points.
Table C-3 Advantages and Disadvantages of Wired Detection of Rogue Access Points
Advantages Disadvantages
• It is easier to monitor networks on a more • This method can miss some rogue access points.
“real-time” basis. • Most of the software is immature or not specifically written to
• It uses automated scripts, which require less detect rogue access points.
manpower to operate. • This method may create a lot of false positives on intrusion
• This method makes it possible to survey detection systems and personal firewalls.
remote sites.
MAC address monitoring tools rely on detecting rogue access points by looking for a known MAC address, or by
cataloging all authorized MAC addresses in the network and looking for new ones. The latter approach has the
advantage of alerting when an unauthorized non-access point device, such as an unauthorized laptop, is connected
to the network. This approach also has more of a problem with false positives. Known MAC address monitoring
tools are listed in Table C-4.
Arpwatch • http://www-nrg.ee.lbl.gov/
Arpwatch monitors Ethernet activity and keeps a database of Ethernet or IP address pairings;
it also reports certain changes via e-mail.
These tools can be utilized to fingerprint an OS, such as the OS on an access point. OS fingerprinting works by
observing particular characteristics of individual OSs such as the way they respond to TCP packets with obscure TCP
flags and options enabled. OS fingerprinting tools are capable of correctly identifying some access points, but the
extent of matching the OS to rogue access points is dependent on the tool having a fingerprint match for the access
point vendor OS. Table C-5 lists the known fingerprinting tools.
NMAP • http://www.insecure.org/nmap/index.html
• http://www.insecure.org/nmap/nmap-fingerprinting-article.html
A very well-known, popular, and respected tool; NMAP is unproven as a rogue access point
detection tool, but it may be useful in conjunction with other rogue access point detection
techniques. It generates a lot of alerts in intrusion detection and personal firewall systems.
Xprobe • http://www.sys-security.com/html/projects/X.html
Xprobe 1 combines various remote active OS fingerprinting methods using the Internet
Control Message Protocol (ICMP), which was discovered during the “ICMP Usage in
Scanning” research project, into a simple, fast, efficient, and powerful way to detect an
underlying OS that a targeted host is using.
Xprobe 2 is an active OS fingerprinting tool with a different approach to OS fingerprinting; it
relies on fuzzy signature matching, probabilistic guesses, multiple matches simultaneously,
and a signature database.
Xprobe is unproven as a rogue access point detection tool, but it may be useful in conjunction
with other rogue access point detection techniques. It generates a lot of alerts in intrusion
detection and personal firewall systems.
SNMP is not thought to be a very effective way to detect rogue access points. Most rogue access points probably
would not have SNMP enabled, and even if they did, SNMP community strings would probably be unknown. If an
SNMP tool is required for rogue access point detection, existing enterprise management packages have the capability
to do IP and SNMP discovery.
IT security administrators can also detect unauthorized WLAN activity by physically observing the work
environment. IT security administrators should be alert for the following:
• Unauthorized WLAN access points in visible locations
• Employees using WLAN access in a location where WLAN access should not be available
• Warchalk symbols denoting WLAN availability (see http://www.warchalking.org/ for more information)
In summary, individually each current rogue access point prevention or detection method is limited. It is
recommended that network designers choose a combination of prevention and detection tools for their network. As
discussed, each prevention or detection tool presents a different capability and can be used in combination with
another tool in order to provide the network designer a comprehensive acceptable rogue access point prevention and
detection toolset.
Wireless Analyzers
Airmagnet • www.airmagnet.com
A commercial product, Airmagnet is a full-featured WLAN site-survey tool that runs on a
Compaq iPaq.
Boingo • www.boingo.com
Boingo is free software that can be downloaded from the Internet; it searches all
available networks, and lets you know when you are in the range of a high-speed service
signal (or tells you where to find the closest one).
Netstumbler • http://www.netstumbler.org/
Very popular and well known, Netstumbler is free software that can be downloaded
from the Internet; it detects WLAN access points and displays information about them.
Sniffer • www.sniffer.com
This professional wireless analyzer could possibly be used to help look for rogue
APaccess points by defining filters to look for beacons, but to exclude authorized SSIDs,
or by defining filters to look for the MAC OUIs of known APaccess point vendors.
Wildpackets • http://www.wildpackets.com/products/airopeek
This professional wireless analyzer could possibly be used to help look for rogue access
points by defining filters to look for beacons, but to exclude authorized SSIDs, or by
defining filters to look for the MAC OUIs of known access point vendors.
Observer • http://www.networkinstruments.com/
This tool could possibly be used to help look for rogue access points by defining filters
to look for beacons, but to exclude authorized SSIDs, or by defining filters to look for the
MAC OUIs of known access point vendors.
Wellenreiter • http://www.remote-exploit.org/
Similar to Netstumbler but less popular and not as well known, Wellenreiter detects
WLAN access points and displays information about them.
Kizmet • http://www.kismetwireless.net/
Kismet is an open source wireless sniffer that could possibly be used to help look for
rogue access points by defining filters to look for beacons, but to exclude authorized
SSIDs.
dachb0den • http://www.dachb0den.com/projects/bsd-airtools.html
This tool, which is not well known, seems to be a combination of Netstumbler and
Airsnort functionality.
Hornet • http://www.bvsystems.com/Products/WLAN/Hornet/hornet.htm
This dedicated hardware looks for a list of access point MAC addresses that have been
configured and downloaded from a PC. It does not seem to do anything that a WLAN
sniffer cannot do.
Table C-7 provides a partial list of MAC OUIs used by access point vendors. This table was obtained from the aptools
site at aptools.sourceforge.net.
3Com 0001.03|0004.76|0050.da|0800.02
Addtron 0040.33|0090.d1
Apple 0030.65
Atmel 0004.25
BreezeNet 0010.e7
Camtec 0000.ff
Compaq 0050.8b
D-Link 0005.5d|0040.05|0090.4b
Intel 0002.b3
Linksys 0003.2f|0004.5a
Lucent 0002.2d|0060.1d|0202.2d
Nokia 00e0.03
Samsung 0000.f0|0002.78
SMC 00e0.29|0090.d1
SOHOware 0080.c6
Sony 0800.46
Symbol 00a0.f8|00a0.0f
Z-Com 0060.b3
Zoom 0040.36
The following sections describe in detail items that should be considered when deploying services in order to secure
wireless LANs (WLANs). Note that in the remote, small, and medium network designs, high availability is not
provided in the Cisco SAFE Blueprint for network security wired network, so it is not expected to be present for
wireless.
RADIUS
• Requests per second—The Remote Access Dial-In User Service (RADIUS) server hardware and software must be
able to accommodate the projected number of new RADIUS requests per second that will be offered by
introducing WLANs. If the RADUIS servers are overburdened, wireless access point and VPN gateways will not
be able to authenticate users, denying wireless users connectivity to the corporate network. Also, it should be
noted that if the network designer elects to use a back-end database for user authentication, the back-end
database must also be designed to accommodate the projected number of user authentication requests per second
that will be offered by introducing WLANs.
• Redundant server deployment—Multiple RADIUS servers should be deployed in order to give the authenticating
device (wireless access point or VPN gateway) redundant options for servicing authentication requests. Network
designers should also group the authenticating devices to alternate the listing of primary and secondary RADIUS
servers. This setup accomplishes two goals: it limits the failure domain in the case of a server failure and also
enables each RADIUS server to scale more effectively.
• User management—RADIUS servers need to provide high-availability access to the user database required for
user authentication. Network designers should consider implementing servers that synchronize data if the user
database is to be stored locally. This setup allows a single point of administration and eliminates the possibility
IP Security Protocol
• Connections per second—The VPN gateway hardware and software must be able to accommodate the projected
number of new IPsec connections per second that will be offered by introducing WLANs.
• Encryption throughput—The VPN gateway hardware and software must be able to accommodate the projected
encryption throughput that will be offered by introducing WLANs. VPN gateways work harder to encrypt several
smaller packets than one larger packet, causing lower encryption throughput numbers for the VPN gateway. It is
important that network designers understand the packet size distribution of their wired networks in order to
properly size the VPN gateway for the wireless network environment.
• Simultaneous IPsec sessions—The VPN gateway hardware and software must be able to accommodate the
projected number of simultaneous IPsec sessions that will be offered by introducing WLANs. VPN gateways are
designed to handle a finite limit of simultaneous IPsec sessions.
Failure to design the IPsec environment with these considerations in mind will cause the wireless users to be unable
to access the corporate network, or when they do so, performance will be severely degraded. VPN vendors have
addressed the previous three items by introducing proprietary clustering technologies. The clustering technologies
load balance new IPsec connections to the least-loaded VPN gateway in order to give the new IPsec connection the
best possible service.
More in-depth information on designing IPsec networks can be found in “SAFE VPN: IPsec Virtual Private Networks
in Depth.”
References
http://www.cisco.com/warp/public/cc/so/cuso/epso/sqfr/safe_wp.htm
SAFE: Extending the Security Blueprint to Small, Midsize, and Remote-User Networks:
http://www.cisco.com/warp/public/cc/so/cuso/epso/sqfr/safes_wp.htm
http://www.cisco.com/warp/public/cc/so/cuso/epso/sqfr/safev_wp.htm
http://www.cisco.com/warp/public/cc/so/cuso/epso/sqfr/snam_wp.htm
http://www.cisco.com/warp/public/cc/so/cuso/epso/sqfr/scdam_wp.htm
http://www.isaac.cs.berkeley.edu/isaac/wep-faq.html
http://www.cs.umd.edu/~waa/wireless.pdf
http://www.cs.umd.edu/~waa/class-pubs/rc4_ksaproc.ps
Using the Fluhrer, Mantin, and Shamir Attack to Break Wired Equivalent Privacy (WEP):
http://www.cs.rice.edu/~astubble/wep/wep_attack.pdf
AirSnort:
http://airsnort.sourceforge.net/
http://www.rsasecurity.com/products/securid/
Acknowledgments
The authors would like to publicly thank all the individuals who contributed to this extension of the SAFE blueprint
and the writing of this document. Certainly, the successful completion of this document would not have been possible
without the valuable input and review feedback from all of the Cisco employees both in corporate headquarters
and in the field. Several individuals provided extra effort in either the review stage or lab validation. The core of this
group included Greg Abelar, Andy Balinsky, Brian Cox, Roland Saville, and Ido Dubrawsky. Thank you all for your
special effort.
1. Borisov et al., “Security of the WEP Algorithm”
Cisco Systems has more than 200 offices in the following countries and regions. Addresses, phone numbers, and fax numbers are listed on the
Cisco Web site at www.cisco.com/go/offices
Argentina • Australia • Austria • Belgium • Brazil • Bulgaria • Canada • Chile • China PRC • Colombia • Costa Rica • Croatia
Czech Republic • Denmark • Dubai, UAE • Finland • France • Germany • Greece • Hong Kong SAR • Hungary • India • Indonesia • Ireland
Israel • Italy • Japan • Korea • Luxembourg • Malaysia • Mexico • The Netherlands • New Zealand • Norway • Peru • Philippines • Poland
Portugal • Puerto Rico • Romania • Russia • Saudi Arabia • Scotland • Singapore • Slovakia • Slovenia • South Africa • Spain • Sweden
S w i t z e r l a n d • Ta i w a n • T h a i l a n d • Tu r k e y • U k r a i n e • U n i t e d K i n g d o m • U n i t e d S t a t e s • Ve n e z u e l a • Vi e t n a m • Z i m b a b w e
All contents are Copyright © 1992–2003 Cisco Systems, Inc. All rights reserved. Aironet, Catalyst, Cisco, Cisco IOS, Cisco Systems, and the Cisco Systems logo are registered trademarks of Cisco Systems, Inc. and/or its
affiliates in the U.S. and certain other countries.
All other trademarks mentioned in this document or Web site are the property of their respective owners. The use of the word partner does not imply a partnership relationship between Cisco and any other company.
(0303R) BU/LW4784 0703