Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

5G slicing and management tmf contribution

•
•
Saurabh Verma
Saurabh Verma

- The document discusses TM Forum's work on 5G network slicing, including requirements, use cases, and business models. - It describes two deployment scenarios: a single slice provider model with one provider spanning access, backhaul, and core networks; and a multi-slice provider model with the end-to-end slice spanning multiple providers. - The key aspects covered are the network slice lifecycle including creation, operations, modification, and termination as well as the roles of 5G OSS/BSS, orchestration, and assurance functions.

Related slideshows

O-RAN 5g high level network design
O-RAN 5g high level network designO-RAN 5g high level network design
O-RAN 5g high level network design
 
5G Service Assurance and Orchestration
5G Service Assurance and Orchestration5G Service Assurance and Orchestration
5G Service Assurance and Orchestration
 
Advanced: Private Networks & 5G Non-Public Networks
Advanced: Private Networks & 5G Non-Public NetworksAdvanced: Private Networks & 5G Non-Public Networks
Advanced: Private Networks & 5G Non-Public Networks
 
1 of 29
Download to read offline
5G Slicing and Management in the TM Forum TM Forum ZOOM Project Team April 26, 2018
TM Forum 5G Work Focus • Learning from related Catalyst (Proof of Concept) Projects since 2016 - With SP deployment scenarios to inform the APIs and modelling development (see supporting slides) • Intent Based Management - Intent against OLA/SLA targets for each tenant instance is required for managing the volume and velocity of 5G services & resources • Closed Control Loop (CCL) - Intent based management needs to support by layered orchestration and closed control loops (which does not imply multiple orchestrator implementations) • Hybrid Infrastructure Management - PNF and VNF operations use platform based approach with CCL and policy mgmt.
5G Essentials: Separation of Concerns • Horizontal Management Operations: Intent based management and cloud native implementation (IaaS PaaS and SaaS) moves operations towards horizontal management of - Service, Resource, Infrastructure /VIM Functions - linked by intent based abstraction using exposed service APIs and agreed SLAs for individual tenants. - 5G Domains – Access/ Backhaul/ Core network 5G ‘Sub Slices’ all modelled using intent based management to support multiple and changing 5G business models i.e. no static relationships between domains but defined exposed services - Intent-based approach allows for easier migration among generations of radio technologies and also support non-radio transport, i.e., the technology-specifics are not embedded in the exposed Service API . • For agility, change management and model driven approaches: Platforms like ONAP should explicitly separate / decouple individual actors’ needs: - The Service View exposed at NBI ExtAPI (Service User actor) - The virtualization and onboarding of Software Assets including NFV (Operational Management actors e.g. SDC) - Management of policies including Security (Enterprise management actors e.g. SDC) - Example anti patterns: Exposing deployment flavors in the NBI ExtAPI Service views – couples implementation to the exposed Service definition making onward change management challenging • Open APIs need to support change by rendering payload dynamically from catalog /repository - 5G slicing needs flexibility to support new features without re-coding API - Early APIs statically bind Payload /DM to API implementation requiring re-coding rather than configuration to accommodate change • TOSCA metadata needs to support full commercial and operational onboarding (e.g., testing, metrics, licensing) , not solely instantiation of Software Images i.e. multiple actors in a lifecycle from supplier to deployed images. - Affects structure and actor usage of ONAP TOSCA templates
5G Essentials: Slicing Model Considerations • Expose technology neutral connectivity models for e2e Service (CFS) and exposed Resource Services (RFS) - Decouple exposed service from internal technical implementation (vendor and technology specific aspects) - Achieved by use of Connectivity Model using exposed (simple) service topology, flows and features - Allows hybrid 5G legacy for 5G slicing capabilities (decouples connectivity models from technology evolution - TM Forum solution defined in TR 255 Connectivity Model ( requirement and use case) covering: • Features constraints and topology models linked to GB 922 Logical Resource Models (Eases integration with Information framework based OSS /BSS e.g. SDC) • The connectivity model is also supported in an intent-based API. • Mapping /integration with detailed resource models and topology e.g. 5G (managed independently by Operational Management e.g. SDC recipes and templates) • End point modelling for aggregation and concatenation of subnetworks using different internal technologies (multi-SDO discussion )
TM Forum Key Assets and Relevance to ONAP (1 of 2) TM Forum Work Activities Relevance to ONAP Requirements/ Use Cases IG1152 Dynamic Network Slices Management and Business Models R17.0.1 • Network slice consists of multiple domains (i.e. Access/Backhaul/Core) • Network slice provided by multiple service providers through APIs • Definitions of actors for 5G e2e management • SDO alignment achieved with Use Case w/ NGMN E2e Arch. , Virtualization w/ ETSI-NFV, 3GPP 23.799 TR229 ZOOM/NFV User Stories Suite R17.0.0 Part A section 4 • Defines actor for Slice Orchestration Policy and Governance • Network Slicing User Stories: Slice Orchestration in Multi-Service Provider with a single Slice Partner; Multi-Service Provider in B2B2C Model; Federated Orchestration Catalyst 5G Intelligent Service Planning and Optimization • Specifically looked at opportunities for integration with VES, DCAE, to achieve E2E Service Assurance 5G Intelligent Service Operations (WIP) • Using ONAP Policy Manager to control Service Orchestration run time behavior Blade Runner (WIP) • 5G enabled AR scenario • Service Assurance/SLA flow through ONAP modules: SDC, SO, SDNC, AAI, External API and VID • TOSCA template for service assurance across CFS and RFS layers
TM Forum Key Assets and Relevance to ONAP (2 of 2) TM Forum Work Activities Relevance to ONAP Intent Based Management • IG1128 Dynamic Control Architecture for Managing a Virtualized Eco-System R16.0.1 • IG1139 Business Rationale and Technical Overview for Orchestration and Autonomic Control Loops R16.0.1 • TR262 Management Platform Blueprint and Application to Hybrid Infrastructure R17.5.1 • IG1161 Overview: Agile Intent-based Resource Management in Hybrid Environments R17.5.1 • ONAP 5G architecture, SO, Closed control loop, Separation of management concerns amongst different actors Modeling • TR255 Connectivity Patterns for Virtualization Management R17.5.1 • GB922 Logical and Compound Resource Computing and Software R17.5.1 supports digital services, multi-tenancy offerings and virtualization/ cloudification by the introduction of the Resource Function concept. • TR275 Core Networking Resources Business Entities R17.5.1 Proposal on federation of parts of the TM Forum Information Framework R17.5 and ONF CIM v1.3 Describes federation between the two models and specific interconnections. This allows Information framework to incorporate transport network models . • SDC, Intent based Modelling of run time configurable and programmable network slicing. • Provides connectivity model with features to abstract /hide internal network detail needs for intent based management..
Next Steps • Identify key areas where ONAP can leverage 5G modeling and API experiences from the TM Forum - Coordinate deeper dive discussions • Identify & prioritize concrete joint Casablanca deliverables - May require multi-SDO coordination
CONFIDENTIAL © 2018 TM Forum | 8 Supporting Materials 8
CONFIDENTIAL © 2018 TM Forum | 9 IG1152 Dynamic Network Slices Management and Business Models
CONFIDENTIAL © 2018 TM Forum | 10  This TM Forum 5G slice approach is Business Scenario and User Story/Use Case-driven with the objective to derive Top 13 slice requirements to pave the way for identifying and demonstrating the value creation for the players interacting in the 5G ecosystem  Targeting relevant technical drivers and overcoming main challenges are key to drive the standardization phases and Products & Solutions maturity and sustainability  Operational 5G System (or Network Slice Platform) as a playground is major piece to be built by the Catalyst partners to assess identified Technical and Business scenarios  The Catalyst cover a small scope and small size then grows as 5G standard matures  3GPP Architecture options considered: Standalone (SA) and Non-Standalone (NSA)  Business scenarios are limited to only two use cases with extreme requirements (eMBB and URLCC). The 5G Catalysts cover (URLCC, eMBB, m IoT, V2X)  Network Slices for those 2 selected use cases could be supported under 2 Deployment Scenarios  A single Slice Provider model where E2E Network Slice spans the Slice Provider’s 3 domains (Access, Backhaul, Core Network) under non-roaming model  Multi-Slice Provider model where E2E Network Slice spans Multi-Slice Provider according to governance, agreement and commercial arrangement among those partners (Roaming model could be an example) Public Document IG1152 Dynamic Network Slices Management and Business Models R17.0.1 Key Contributors IG1152 Dynamic Network Slices Management and Business Models
CONFIDENTIAL © 2018 TM Forum | 11 E2E Network Life Cycle Management: Architecture and Processes (Holistic view: Customer & SP)  Dynamic Creation  Slice Registration  Operations  Orchestration  Autonomic & Cognitive Assurance  Modification  Termination  Retirement 5G OSS EM, VNFM, VIM, SDN Controller empowered by Autonomic & Cognitive capabilities yc Data Analytics 5G BSS Slice Provider’s perspectiveCustomer’s perspective Processes Network Slice Dynamic Design Tool-Chain E2E IoT Network Slice Instance E2E eMBB Network Slice Instance 5G Access + LTE (SON) 5G Core + EPC (SON-like / Autonomics) Backhaul (SON-like / Autonomics) IoTIoT eMBBeMBB  Customer Onboarding  Subscription  Network Slice (Self) Ordering  Agile Delivery  Dynamic change request  Charging  Billing  Settlement  Editors for Network Slice (NFs)  Policy Generators (from Network Objectives / Goals)  Validators against conflicts (Goals conflict resolution, Policy conflict resolution, Bad NFs chaining resolution)
CONFIDENTIAL © 2018 TM Forum | 12 Network Slice Deployment Scenario 1: in Single Slice Provider (Multi-Domain model) Customers / Tenants ’ Network Slice Delivery Points IoTNFs IoTNFs IoTNFs 5 G O S S / B S S RAN eMBB Network Slice Core eMBB Network SliceBackhaul eMBB Network Slice Backhaul Domain 5G Core Network Domain RAN URLLC Network Slice Backhaul IoT Network Slice Core IoT Network Slice DynamicNetworkSlice Design Policy&Governance SP’s Portal 5G RAN Domain Backhaul Network Slice Platform 5G Core Network Slice Platform5G RAN Network Slice Platform eMBBNFs eMBBNFs eMBBNFs (Self) OrderAPI Delivery C&BAPI Customers’ 2 E2E Network Slice Instances APISlice Policy API API API API API RAN User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF 5G RAN NFs Catalog NFs States Backhaul User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF X-haul NFs Catalog NFs States Core User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF 5G Core NFs Catalog NFs States Domain Domain Domain SingleProvider
CONFIDENTIAL © 2018 TM Forum | 13 Network Slice Deployment Scenario: 2: in Multi-Slice Provider model Customers/Tenants ’ Network Slice Delivery Points IoTNFs IoTNFs IoTNFs 5 G O S S / B S S SP’s eMBB Network Slice SP2’s eMBB Network SliceSP1’s eMBB Network Slice Slice Provider 1 5G System Slice Provider 2 5G System SP’s IoT Network Slice SP1’s IoT Network Slice SP2’s IoT Network Slice DynamicNetworkSliceDesign Muri-PartnerPolicy& Governance SP’s Portal Service Provider 5G System SP1’s Network Slice Platform SP2’s Network Slice PlatformSP’s Network Slice Platform eMBBNFs eMBBNFs eMBBNFs (Self) OrderAPI Delivery C&BAPI Customers’ 2 E2E Network Slice Instances APISlice Policy API API API API API SP’s User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF SP’s NFs Catalog NFs States SP1’s User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF SP1’s NFs Catalog NFs States SP2’s User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF SP2’ NFs Catalog NFs States Service Provider 1 Service Provider 2 Service Provider 3 Multi-Provider
CONFIDENTIAL © 2018 TM Forum | 14 • Network Slice is designed as composition of NFs at CP and UP according to 3GPPSA2 TR 23 79 – Control Plane NFs – User Plane NFs • Each of the those Resources Capabilities (RFs) are stored in a catalog. We are using TM Forum terminology with those resource capabilities (RF: Resource Function) Network Slice structure options: Network Functions composing a Network Slice instance
CONFIDENTIAL © 2018 TM Forum | 15 Catalyst: 5G Service Operations – Real Time Service Assurance
CONFIDENTIAL © 2018 TM Forum | 16 5G Challenges 16 ▪ 5G Service Management and Orchestration ― Many POCs and research on 5G radio aspects but very little on operations and management. ― No clear agreement on what needs to be managed or how the management should be organised ― Expectation of radical change in how 5G operations is organised and the OSS IT systems to support them ― Impact of 5G on service orchestration for slices not fully explored ― Create, manage and automate end-to-end service- specific network slices that cross legacy, SDN, mobile and fixed networks ― SDN/NFV operations platforms need enhancement to deliver new services ― Full automation from fulfilment to assurance processes to support billions of IoT connections ▪ New 5G services ― Assigning specific network characteristics for each type of service across the network. ― Defining services with specific attributes such as latency, bandwidth etc by creation of network slices ― Assuring and guaranteeing service performance and SLAs per network slice ▪ 5G Network Slicing ― Currently on a service class basis with services mapped into 3-4 network slices manually ― 5G based IoT services (with different attributes) need to be treated individually for better network utilization ― Tens of thousands of slices require the entire process to be automated
CONFIDENTIAL © 2018 TM Forum | 17 Catalyst Scope ▪ Focus on two 5G based IoT use cases with the goal of: ― Composing repeatable services for industry verticals in hours ― Zero friction integration into OSS /BSS Operations Center of the Future ― Dynamically and autonomously fulfilling and assuring numerous concurrent slice instances (in parallel) with total isolation and security to meet carrier-grade standards. ▪ Explore the impact of 5G networks and network slicing on operations in creating and automating many parallel end-to-end network slices that cross multiple network types to: ― Explore technology readiness and close potential gaps in 5G operations ― Enable CSP to automate the creation and modification of specific network slices for new IoT services with specific SLAs and characteristics ― Leverage open source to address the needs of IoT and to accelerate innovation uRLLC (Ultra-reliable and low latency communication) uMTC (Ultra-reliable machine-type communication) Connected Factories Connected Cars
CONFIDENTIAL © 2018 TM Forum | 18 5G Operation – Management platforms and flow Service AssuranceService Orchestrator 2. Network Slice On-boarding 3. . Network Slice Inventory/Lifecycle 4. Performance Collection and Thresholding TMF 628 5G Network Slice Closed-loop Assurance VNF Event Stream (ONAP) Metrics Simulation Platform Capability 5G Radio Simulation Platform Capability 5. uMTC Slice Type Operations 5. Slice Scaling request (Scaling API evolving towards Entity Provisioning API) 5. uRLLC Slice Type SLA Operations 1. Slices creation and fulfillment 18
CONFIDENTIAL © 2018 TM Forum | 19 Connected Factory – uRLLC slice operations Smart Factory Applications Platform 5G RAN Node 5G Cell 5G RAN Slice Smart Factory Island uRLLC uMBB uRLLC Sensor Slice uRLLC Robots Slice uMBB 5G Cell5G Cell 5G Common CPNF Data Network Core Network Network Slice Flows (NG2/NG3/NG6) 5G Slice UPNF uRLLC-Slice- Sensors uRLLC-Slice- RobotCams uMBB-Slice Latency <10ms SLA <10ms SLA <50ms DataRate-uplink per Device 200 Kbps CBR per Device 50 Mbps CBR per Device 5 Mbps VBR DataRate-downlink per Device 100 Kbps CBR per Device 200 Kbps CBR per Device 5 Mbps VBR per Device Reliability < 0.3% < 0.3% < 1% Pro-active SLA and Demand MonitoringDynamic Scaling (vnfcScalingMetric, Slice allocated bandwidth (MHz)) 19
CONFIDENTIAL © 2018 TM Forum | 20 Connected Car- uMTC slice operations 5G RAN Node 5G Cell 5G RAN Slice uMTC (V2N, V2V) Car Companies Connected Car Platforms & apps Proactive Network Monitoring 5G Cell5G Cell uMTC (V2I) uMTC Sensor Slice mBB (Car as a Mobile Hotspot) 5G Common CPNF Data Network Core NetworkNetwork Slice Flows (NG2/NG3/NG6) 5G Slice UPNF uMTC for V2x commnuications mBB-Slice Latency (ms) SLA<10ms SLA <50ms Availability (%) SLA > 99% >95% Reliability (%) SLA > 99.99% >98% Slice Utilization (%) < 75% < 85% User Experienced Data Rate (UL/DL) (Mbps) DL > 5 UL > 1 DL > 30 UL > 5 Proactive Alarming, Correlation Closed Loop AutomationDynamic Scaling (vnfcScalingMetric, Slice allocated bandwidth (MHz)) mBB
CONFIDENTIAL © 2018 TM Forum | 21 ▪ Devised an efficient system to operationalize 5G slices using: ― Closed loop assurance ― Policy-based automation ― Domain/service Orchestration ▪ Tested the system for dynamic and autonomous Fulfillment and Assurance for concurrent 5G Slices ▪ Gained clarity on QoS requirements for 5G service categories: uRLLC and uMTC ▪ Established the role of APIs in the 5G ecosystem: how to leverage open source to address the needs of IoT ▪ Demonstrated how 5G ecosystem integration can be carried out through TM Forum concepts and API operations in: ― TM 628 Performance Management API, ― TR 255 Entity Provisioning API, (aka Resources Function Configuration and Activation API) ― Service Assurance using intent based SLA/OLA approach ▪ Validated the ZOOM TR 262 Hybrid Infrastructure Platform principles and requirements Outcome
CONFIDENTIAL © 2018 TM Forum | 22 5G Connected Factory uRLLC Use Case Latency Data Rate Closed Loop Assurance uRLLC slice type Assurance Policy-based automation (Demand and Congestion) uRLLC slice type Assurance SLA proactive monitoring uRLLC Slice type SLA Conformance Alarming uRLLC slice type Assurance On-boarding uRLLC slice-type scaling orchestration uRLLC slice-type on- boarding / instantiation Resource (re)allocation Scaling request
CONFIDENTIAL © 2018 TM Forum | 23 5G Connected Cars uMTC Use Case Closed Loop Assurance Rule-based Automated activation of scale-out and scale-in commands per monitored slice(s) uMTC (V2X) Integrated monitoring of contextual car data uMTC slice type Assurance On-boarding uMTC slice type On- boarding / Instantiation Scaling request Application of smart thresholds and correlation to proactively monitor SLA conformance uMTC slice type Orchestration tasks
CONFIDENTIAL © 2018 TM Forum | 24 Consumer, operational and Network scenarios
CONFIDENTIAL © 2018 TM Forum | 25 • The 5G Intelligent Service Operations is demonstrating the usage of ONAP Policy management integrated with external OSS/BSS systems as part of an operational closed-loop use case. Key Contributors 5G Intelligent Service Operations Catalyst – ONAP Policies • Business-Service Orchestration policies • Service-slice policies • Centralized and distributed policy management
CONFIDENTIAL © 2018 TM Forum | 26 ONAP Solution for Blade Runner Catalyst
CONFIDENTIAL © 2018 TM Forum | 27 Blade Runner Catalyst – ONAP Solution Overview ONAP SDC ONAP SO SDNC VIM (OpenStack - Ocata) ONAP AAI EXFO Probe (vVerifier)Riverbed SD WAN Fortinet Firewall Domain Orchestrator Video Server External API Plug In Service Catalog (633) Service Order (641) Service Inventory (638) VNFs /RFS: EXFO vVerifier, Riverbed SDWAN, Video App, Fortinet Riverbed Master Orchestrator RESTconf (YANG) GET GETPOST REST (Heat) • Onboarding VNFs on ONAP • Instantiating VNFs on Openstack Via ONAP • Scale out Network Service Service Designer / Orchestrator/Inventory RIFT.io CFS: Internet Reference to RFS VIDONAP Portal CSAR REST/JSON
CONFIDENTIAL © 2018 TM Forum | 28 Domain Specific Resources Facing Service Management ONAP Domain Orchestrator End to End Customer Facing Service Management Service Orchestrator Technology Specific Resource Management End to End Customer Facing Service Management VIM SD-WAN, Firewall, Probes Domain Specific Resources Facing Service Management ONAP Domain Orchestrator Technology Specific Resource Management VIM SD-WAN, AR, Firewall, Probes Product & Offer ManagementProduct & Offer Management Commercial Product CatalogueCommercial Product Catalogue Service OrchestratorService Assurance Service Assurance Blade Runner: Open Digital Architecture | Partners View Customer & Partner Management CRM Wholesale Billing Customer & Partner Management CRM Wholesale Billing Proprietary integration Interlude Riverbed Domain Orchestrator
CONFIDENTIAL © 2018 TM Forum | 29 ONAP Domain Orchestrator Service Orchestrator VIM SD-WAN, Firewall, Probes ONAP Domain Orchestrator VIM SD-WAN, AR, Firewall, Probes Service OrchestratorService Assurance Service Assurance Blade Runner: Systems Integration on a page CRM Wholesale Billing Proprietary integration Interlude ServiceCatalogueAPI Service Problem Management API?? ServiceCatalogueAPI ServiceOrderAPI Service Catalogue Riverbed Domain Orchestrator Service Inventory (Topology) Probes Collector Controller (SteelConnect) ServiceCatalogueAPI NETCONF NETCONF HEAT HEAT Probes Collector HEAT Service Catalogue Service Inventory (Topology) ServiceOrderAPI BearingPoint Cloud AWS Vodafone - US East (Ohio) AWS Vodafone - EU (London) Azure Infosys AWS Riverbed ServiceInventoryAPI ServiceInventoryAPI AR Client Trigger corrective action Service Order API (MEF Interlude Rf. Point)

More Related Content

5G slicing and management tmf contribution

  • 1. 5G Slicing and Management in the TM Forum TM Forum ZOOM Project Team April 26, 2018
  • 2. TM Forum 5G Work Focus • Learning from related Catalyst (Proof of Concept) Projects since 2016 - With SP deployment scenarios to inform the APIs and modelling development (see supporting slides) • Intent Based Management - Intent against OLA/SLA targets for each tenant instance is required for managing the volume and velocity of 5G services & resources • Closed Control Loop (CCL) - Intent based management needs to support by layered orchestration and closed control loops (which does not imply multiple orchestrator implementations) • Hybrid Infrastructure Management - PNF and VNF operations use platform based approach with CCL and policy mgmt.
  • 3. 5G Essentials: Separation of Concerns • Horizontal Management Operations: Intent based management and cloud native implementation (IaaS PaaS and SaaS) moves operations towards horizontal management of - Service, Resource, Infrastructure /VIM Functions - linked by intent based abstraction using exposed service APIs and agreed SLAs for individual tenants. - 5G Domains – Access/ Backhaul/ Core network 5G ‘Sub Slices’ all modelled using intent based management to support multiple and changing 5G business models i.e. no static relationships between domains but defined exposed services - Intent-based approach allows for easier migration among generations of radio technologies and also support non-radio transport, i.e., the technology-specifics are not embedded in the exposed Service API . • For agility, change management and model driven approaches: Platforms like ONAP should explicitly separate / decouple individual actors’ needs: - The Service View exposed at NBI ExtAPI (Service User actor) - The virtualization and onboarding of Software Assets including NFV (Operational Management actors e.g. SDC) - Management of policies including Security (Enterprise management actors e.g. SDC) - Example anti patterns: Exposing deployment flavors in the NBI ExtAPI Service views – couples implementation to the exposed Service definition making onward change management challenging • Open APIs need to support change by rendering payload dynamically from catalog /repository - 5G slicing needs flexibility to support new features without re-coding API - Early APIs statically bind Payload /DM to API implementation requiring re-coding rather than configuration to accommodate change • TOSCA metadata needs to support full commercial and operational onboarding (e.g., testing, metrics, licensing) , not solely instantiation of Software Images i.e. multiple actors in a lifecycle from supplier to deployed images. - Affects structure and actor usage of ONAP TOSCA templates
  • 4. 5G Essentials: Slicing Model Considerations • Expose technology neutral connectivity models for e2e Service (CFS) and exposed Resource Services (RFS) - Decouple exposed service from internal technical implementation (vendor and technology specific aspects) - Achieved by use of Connectivity Model using exposed (simple) service topology, flows and features - Allows hybrid 5G legacy for 5G slicing capabilities (decouples connectivity models from technology evolution - TM Forum solution defined in TR 255 Connectivity Model ( requirement and use case) covering: • Features constraints and topology models linked to GB 922 Logical Resource Models (Eases integration with Information framework based OSS /BSS e.g. SDC) • The connectivity model is also supported in an intent-based API. • Mapping /integration with detailed resource models and topology e.g. 5G (managed independently by Operational Management e.g. SDC recipes and templates) • End point modelling for aggregation and concatenation of subnetworks using different internal technologies (multi-SDO discussion )
  • 5. TM Forum Key Assets and Relevance to ONAP (1 of 2) TM Forum Work Activities Relevance to ONAP Requirements/ Use Cases IG1152 Dynamic Network Slices Management and Business Models R17.0.1 • Network slice consists of multiple domains (i.e. Access/Backhaul/Core) • Network slice provided by multiple service providers through APIs • Definitions of actors for 5G e2e management • SDO alignment achieved with Use Case w/ NGMN E2e Arch. , Virtualization w/ ETSI-NFV, 3GPP 23.799 TR229 ZOOM/NFV User Stories Suite R17.0.0 Part A section 4 • Defines actor for Slice Orchestration Policy and Governance • Network Slicing User Stories: Slice Orchestration in Multi-Service Provider with a single Slice Partner; Multi-Service Provider in B2B2C Model; Federated Orchestration Catalyst 5G Intelligent Service Planning and Optimization • Specifically looked at opportunities for integration with VES, DCAE, to achieve E2E Service Assurance 5G Intelligent Service Operations (WIP) • Using ONAP Policy Manager to control Service Orchestration run time behavior Blade Runner (WIP) • 5G enabled AR scenario • Service Assurance/SLA flow through ONAP modules: SDC, SO, SDNC, AAI, External API and VID • TOSCA template for service assurance across CFS and RFS layers
  • 6. TM Forum Key Assets and Relevance to ONAP (2 of 2) TM Forum Work Activities Relevance to ONAP Intent Based Management • IG1128 Dynamic Control Architecture for Managing a Virtualized Eco-System R16.0.1 • IG1139 Business Rationale and Technical Overview for Orchestration and Autonomic Control Loops R16.0.1 • TR262 Management Platform Blueprint and Application to Hybrid Infrastructure R17.5.1 • IG1161 Overview: Agile Intent-based Resource Management in Hybrid Environments R17.5.1 • ONAP 5G architecture, SO, Closed control loop, Separation of management concerns amongst different actors Modeling • TR255 Connectivity Patterns for Virtualization Management R17.5.1 • GB922 Logical and Compound Resource Computing and Software R17.5.1 supports digital services, multi-tenancy offerings and virtualization/ cloudification by the introduction of the Resource Function concept. • TR275 Core Networking Resources Business Entities R17.5.1 Proposal on federation of parts of the TM Forum Information Framework R17.5 and ONF CIM v1.3 Describes federation between the two models and specific interconnections. This allows Information framework to incorporate transport network models . • SDC, Intent based Modelling of run time configurable and programmable network slicing. • Provides connectivity model with features to abstract /hide internal network detail needs for intent based management..
  • 7. Next Steps • Identify key areas where ONAP can leverage 5G modeling and API experiences from the TM Forum - Coordinate deeper dive discussions • Identify & prioritize concrete joint Casablanca deliverables - May require multi-SDO coordination
  • 8. CONFIDENTIAL © 2018 TM Forum | 8 Supporting Materials 8
  • 9. CONFIDENTIAL © 2018 TM Forum | 9 IG1152 Dynamic Network Slices Management and Business Models
  • 10. CONFIDENTIAL © 2018 TM Forum | 10  This TM Forum 5G slice approach is Business Scenario and User Story/Use Case-driven with the objective to derive Top 13 slice requirements to pave the way for identifying and demonstrating the value creation for the players interacting in the 5G ecosystem  Targeting relevant technical drivers and overcoming main challenges are key to drive the standardization phases and Products & Solutions maturity and sustainability  Operational 5G System (or Network Slice Platform) as a playground is major piece to be built by the Catalyst partners to assess identified Technical and Business scenarios  The Catalyst cover a small scope and small size then grows as 5G standard matures  3GPP Architecture options considered: Standalone (SA) and Non-Standalone (NSA)  Business scenarios are limited to only two use cases with extreme requirements (eMBB and URLCC). The 5G Catalysts cover (URLCC, eMBB, m IoT, V2X)  Network Slices for those 2 selected use cases could be supported under 2 Deployment Scenarios  A single Slice Provider model where E2E Network Slice spans the Slice Provider’s 3 domains (Access, Backhaul, Core Network) under non-roaming model  Multi-Slice Provider model where E2E Network Slice spans Multi-Slice Provider according to governance, agreement and commercial arrangement among those partners (Roaming model could be an example) Public Document IG1152 Dynamic Network Slices Management and Business Models R17.0.1 Key Contributors IG1152 Dynamic Network Slices Management and Business Models
  • 11. CONFIDENTIAL © 2018 TM Forum | 11 E2E Network Life Cycle Management: Architecture and Processes (Holistic view: Customer & SP)  Dynamic Creation  Slice Registration  Operations  Orchestration  Autonomic & Cognitive Assurance  Modification  Termination  Retirement 5G OSS EM, VNFM, VIM, SDN Controller empowered by Autonomic & Cognitive capabilities yc Data Analytics 5G BSS Slice Provider’s perspectiveCustomer’s perspective Processes Network Slice Dynamic Design Tool-Chain E2E IoT Network Slice Instance E2E eMBB Network Slice Instance 5G Access + LTE (SON) 5G Core + EPC (SON-like / Autonomics) Backhaul (SON-like / Autonomics) IoTIoT eMBBeMBB  Customer Onboarding  Subscription  Network Slice (Self) Ordering  Agile Delivery  Dynamic change request  Charging  Billing  Settlement  Editors for Network Slice (NFs)  Policy Generators (from Network Objectives / Goals)  Validators against conflicts (Goals conflict resolution, Policy conflict resolution, Bad NFs chaining resolution)
  • 12. CONFIDENTIAL © 2018 TM Forum | 12 Network Slice Deployment Scenario 1: in Single Slice Provider (Multi-Domain model) Customers / Tenants ’ Network Slice Delivery Points IoTNFs IoTNFs IoTNFs 5 G O S S / B S S RAN eMBB Network Slice Core eMBB Network SliceBackhaul eMBB Network Slice Backhaul Domain 5G Core Network Domain RAN URLLC Network Slice Backhaul IoT Network Slice Core IoT Network Slice DynamicNetworkSlice Design Policy&Governance SP’s Portal 5G RAN Domain Backhaul Network Slice Platform 5G Core Network Slice Platform5G RAN Network Slice Platform eMBBNFs eMBBNFs eMBBNFs (Self) OrderAPI Delivery C&BAPI Customers’ 2 E2E Network Slice Instances APISlice Policy API API API API API RAN User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF 5G RAN NFs Catalog NFs States Backhaul User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF X-haul NFs Catalog NFs States Core User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF 5G Core NFs Catalog NFs States Domain Domain Domain SingleProvider
  • 13. CONFIDENTIAL © 2018 TM Forum | 13 Network Slice Deployment Scenario: 2: in Multi-Slice Provider model Customers/Tenants ’ Network Slice Delivery Points IoTNFs IoTNFs IoTNFs 5 G O S S / B S S SP’s eMBB Network Slice SP2’s eMBB Network SliceSP1’s eMBB Network Slice Slice Provider 1 5G System Slice Provider 2 5G System SP’s IoT Network Slice SP1’s IoT Network Slice SP2’s IoT Network Slice DynamicNetworkSliceDesign Muri-PartnerPolicy& Governance SP’s Portal Service Provider 5G System SP1’s Network Slice Platform SP2’s Network Slice PlatformSP’s Network Slice Platform eMBBNFs eMBBNFs eMBBNFs (Self) OrderAPI Delivery C&BAPI Customers’ 2 E2E Network Slice Instances APISlice Policy API API API API API SP’s User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF SP’s NFs Catalog NFs States SP1’s User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF SP1’s NFs Catalog NFs States SP2’s User Plane NFs UPFs Common Control Plane NFs AMF SMF NSSF SP2’ NFs Catalog NFs States Service Provider 1 Service Provider 2 Service Provider 3 Multi-Provider
  • 14. CONFIDENTIAL © 2018 TM Forum | 14 • Network Slice is designed as composition of NFs at CP and UP according to 3GPPSA2 TR 23 79 – Control Plane NFs – User Plane NFs • Each of the those Resources Capabilities (RFs) are stored in a catalog. We are using TM Forum terminology with those resource capabilities (RF: Resource Function) Network Slice structure options: Network Functions composing a Network Slice instance
  • 15. CONFIDENTIAL © 2018 TM Forum | 15 Catalyst: 5G Service Operations – Real Time Service Assurance
  • 16. CONFIDENTIAL © 2018 TM Forum | 16 5G Challenges 16 ▪ 5G Service Management and Orchestration ― Many POCs and research on 5G radio aspects but very little on operations and management. ― No clear agreement on what needs to be managed or how the management should be organised ― Expectation of radical change in how 5G operations is organised and the OSS IT systems to support them ― Impact of 5G on service orchestration for slices not fully explored ― Create, manage and automate end-to-end service- specific network slices that cross legacy, SDN, mobile and fixed networks ― SDN/NFV operations platforms need enhancement to deliver new services ― Full automation from fulfilment to assurance processes to support billions of IoT connections ▪ New 5G services ― Assigning specific network characteristics for each type of service across the network. ― Defining services with specific attributes such as latency, bandwidth etc by creation of network slices ― Assuring and guaranteeing service performance and SLAs per network slice ▪ 5G Network Slicing ― Currently on a service class basis with services mapped into 3-4 network slices manually ― 5G based IoT services (with different attributes) need to be treated individually for better network utilization ― Tens of thousands of slices require the entire process to be automated
  • 17. CONFIDENTIAL © 2018 TM Forum | 17 Catalyst Scope ▪ Focus on two 5G based IoT use cases with the goal of: ― Composing repeatable services for industry verticals in hours ― Zero friction integration into OSS /BSS Operations Center of the Future ― Dynamically and autonomously fulfilling and assuring numerous concurrent slice instances (in parallel) with total isolation and security to meet carrier-grade standards. ▪ Explore the impact of 5G networks and network slicing on operations in creating and automating many parallel end-to-end network slices that cross multiple network types to: ― Explore technology readiness and close potential gaps in 5G operations ― Enable CSP to automate the creation and modification of specific network slices for new IoT services with specific SLAs and characteristics ― Leverage open source to address the needs of IoT and to accelerate innovation uRLLC (Ultra-reliable and low latency communication) uMTC (Ultra-reliable machine-type communication) Connected Factories Connected Cars
  • 18. CONFIDENTIAL © 2018 TM Forum | 18 5G Operation – Management platforms and flow Service AssuranceService Orchestrator 2. Network Slice On-boarding 3. . Network Slice Inventory/Lifecycle 4. Performance Collection and Thresholding TMF 628 5G Network Slice Closed-loop Assurance VNF Event Stream (ONAP) Metrics Simulation Platform Capability 5G Radio Simulation Platform Capability 5. uMTC Slice Type Operations 5. Slice Scaling request (Scaling API evolving towards Entity Provisioning API) 5. uRLLC Slice Type SLA Operations 1. Slices creation and fulfillment 18
  • 19. CONFIDENTIAL © 2018 TM Forum | 19 Connected Factory – uRLLC slice operations Smart Factory Applications Platform 5G RAN Node 5G Cell 5G RAN Slice Smart Factory Island uRLLC uMBB uRLLC Sensor Slice uRLLC Robots Slice uMBB 5G Cell5G Cell 5G Common CPNF Data Network Core Network Network Slice Flows (NG2/NG3/NG6) 5G Slice UPNF uRLLC-Slice- Sensors uRLLC-Slice- RobotCams uMBB-Slice Latency <10ms SLA <10ms SLA <50ms DataRate-uplink per Device 200 Kbps CBR per Device 50 Mbps CBR per Device 5 Mbps VBR DataRate-downlink per Device 100 Kbps CBR per Device 200 Kbps CBR per Device 5 Mbps VBR per Device Reliability < 0.3% < 0.3% < 1% Pro-active SLA and Demand MonitoringDynamic Scaling (vnfcScalingMetric, Slice allocated bandwidth (MHz)) 19
  • 20. CONFIDENTIAL © 2018 TM Forum | 20 Connected Car- uMTC slice operations 5G RAN Node 5G Cell 5G RAN Slice uMTC (V2N, V2V) Car Companies Connected Car Platforms & apps Proactive Network Monitoring 5G Cell5G Cell uMTC (V2I) uMTC Sensor Slice mBB (Car as a Mobile Hotspot) 5G Common CPNF Data Network Core NetworkNetwork Slice Flows (NG2/NG3/NG6) 5G Slice UPNF uMTC for V2x commnuications mBB-Slice Latency (ms) SLA<10ms SLA <50ms Availability (%) SLA > 99% >95% Reliability (%) SLA > 99.99% >98% Slice Utilization (%) < 75% < 85% User Experienced Data Rate (UL/DL) (Mbps) DL > 5 UL > 1 DL > 30 UL > 5 Proactive Alarming, Correlation Closed Loop AutomationDynamic Scaling (vnfcScalingMetric, Slice allocated bandwidth (MHz)) mBB
  • 21. CONFIDENTIAL © 2018 TM Forum | 21 ▪ Devised an efficient system to operationalize 5G slices using: ― Closed loop assurance ― Policy-based automation ― Domain/service Orchestration ▪ Tested the system for dynamic and autonomous Fulfillment and Assurance for concurrent 5G Slices ▪ Gained clarity on QoS requirements for 5G service categories: uRLLC and uMTC ▪ Established the role of APIs in the 5G ecosystem: how to leverage open source to address the needs of IoT ▪ Demonstrated how 5G ecosystem integration can be carried out through TM Forum concepts and API operations in: ― TM 628 Performance Management API, ― TR 255 Entity Provisioning API, (aka Resources Function Configuration and Activation API) ― Service Assurance using intent based SLA/OLA approach ▪ Validated the ZOOM TR 262 Hybrid Infrastructure Platform principles and requirements Outcome
  • 22. CONFIDENTIAL © 2018 TM Forum | 22 5G Connected Factory uRLLC Use Case Latency Data Rate Closed Loop Assurance uRLLC slice type Assurance Policy-based automation (Demand and Congestion) uRLLC slice type Assurance SLA proactive monitoring uRLLC Slice type SLA Conformance Alarming uRLLC slice type Assurance On-boarding uRLLC slice-type scaling orchestration uRLLC slice-type on- boarding / instantiation Resource (re)allocation Scaling request
  • 23. CONFIDENTIAL © 2018 TM Forum | 23 5G Connected Cars uMTC Use Case Closed Loop Assurance Rule-based Automated activation of scale-out and scale-in commands per monitored slice(s) uMTC (V2X) Integrated monitoring of contextual car data uMTC slice type Assurance On-boarding uMTC slice type On- boarding / Instantiation Scaling request Application of smart thresholds and correlation to proactively monitor SLA conformance uMTC slice type Orchestration tasks
  • 24. CONFIDENTIAL © 2018 TM Forum | 24 Consumer, operational and Network scenarios
  • 25. CONFIDENTIAL © 2018 TM Forum | 25 • The 5G Intelligent Service Operations is demonstrating the usage of ONAP Policy management integrated with external OSS/BSS systems as part of an operational closed-loop use case. Key Contributors 5G Intelligent Service Operations Catalyst – ONAP Policies • Business-Service Orchestration policies • Service-slice policies • Centralized and distributed policy management
  • 26. CONFIDENTIAL © 2018 TM Forum | 26 ONAP Solution for Blade Runner Catalyst
  • 27. CONFIDENTIAL © 2018 TM Forum | 27 Blade Runner Catalyst – ONAP Solution Overview ONAP SDC ONAP SO SDNC VIM (OpenStack - Ocata) ONAP AAI EXFO Probe (vVerifier)Riverbed SD WAN Fortinet Firewall Domain Orchestrator Video Server External API Plug In Service Catalog (633) Service Order (641) Service Inventory (638) VNFs /RFS: EXFO vVerifier, Riverbed SDWAN, Video App, Fortinet Riverbed Master Orchestrator RESTconf (YANG) GET GETPOST REST (Heat) • Onboarding VNFs on ONAP • Instantiating VNFs on Openstack Via ONAP • Scale out Network Service Service Designer / Orchestrator/Inventory RIFT.io CFS: Internet Reference to RFS VIDONAP Portal CSAR REST/JSON
  • 28. CONFIDENTIAL © 2018 TM Forum | 28 Domain Specific Resources Facing Service Management ONAP Domain Orchestrator End to End Customer Facing Service Management Service Orchestrator Technology Specific Resource Management End to End Customer Facing Service Management VIM SD-WAN, Firewall, Probes Domain Specific Resources Facing Service Management ONAP Domain Orchestrator Technology Specific Resource Management VIM SD-WAN, AR, Firewall, Probes Product & Offer ManagementProduct & Offer Management Commercial Product CatalogueCommercial Product Catalogue Service OrchestratorService Assurance Service Assurance Blade Runner: Open Digital Architecture | Partners View Customer & Partner Management CRM Wholesale Billing Customer & Partner Management CRM Wholesale Billing Proprietary integration Interlude Riverbed Domain Orchestrator
  • 29. CONFIDENTIAL © 2018 TM Forum | 29 ONAP Domain Orchestrator Service Orchestrator VIM SD-WAN, Firewall, Probes ONAP Domain Orchestrator VIM SD-WAN, AR, Firewall, Probes Service OrchestratorService Assurance Service Assurance Blade Runner: Systems Integration on a page CRM Wholesale Billing Proprietary integration Interlude ServiceCatalogueAPI Service Problem Management API?? ServiceCatalogueAPI ServiceOrderAPI Service Catalogue Riverbed Domain Orchestrator Service Inventory (Topology) Probes Collector Controller (SteelConnect) ServiceCatalogueAPI NETCONF NETCONF HEAT HEAT Probes Collector HEAT Service Catalogue Service Inventory (Topology) ServiceOrderAPI BearingPoint Cloud AWS Vodafone - US East (Ohio) AWS Vodafone - EU (London) Azure Infosys AWS Riverbed ServiceInventoryAPI ServiceInventoryAPI AR Client Trigger corrective action Service Order API (MEF Interlude Rf. Point)