Enabling Multi-access Edge Computing (MEC) Platform-as-a-Service for Enterprises
- 1. Krishnamurthy Jambur (krishna.j.murthy@intel.com) DCG / NPG / 5GID / SWA
- 2. ACCESS/EDGEDevices/Things Cloud/DatacenterCORE Enterprise/ Factories Home Small Cells licensed & unlicensed Macro Cells Wireless Carrier WiFi Wireline On-premise Equipment Central Offices Regional Central Offices & Data Centers Network Edge Enterprise View of Edge CoSP’s view of Edge CSP’s view of Edge
- 3. Intel® Network Edge Virtualization (NEV) SDK For Multi-access Edge Computing (MEC) Features • Reference Libraries – MEC REST APIs and Optimized MEC data path Routing • S1, SGi (EPC), WiFi traffic and Local Breakout support • Sample Application / MEC Child Application Support • VLAN Tags processing • Wind River optimized telco cloud for MEC and vRAN Deployment • OpenStack components • Standalone deployment support with CentOS • Kubernetes container orchestration • AWS Greengrass and Azure IoT Edge support • Intel® System Studio Suite • Test Suites, Programming Guide, User Guide and App Notes Release 1: Hypervisor Wind River Titanium Server Components (on the Host Platform) MEC Apps MEC Services Release 2: Release 3: WR TS4/5 and CentOS-7 Intel ®Atom™ Processor C2000 Platform Intel® XEON® Platform Software bundle MEC IOT-GW MEC Services Your Apps go here ENB and EPC EmulatorsTelco Cloud NFV infrastructure Optimization and analysis tools Intel, the Intel logo are trademarks of Intel Corporation or its subsidiaries in the U.S. and/or other countries. *Other names and brands may be claimed as the property of others. Release 4: Cloud Connectors Release 5: Cloud Native
- 4. NEV SDK: 4G and 5G Support • Current version of NEV SDK supports Rel. 14 LTE deployments. In LTE deployments are supported on S1u and SGi • For 5G SA the NEV SDK will be updated to support the SA changes for both Data plane and Service based Architecture. Data Network EC Apps EC Platforms EC Apps EC Platforms S1u Deployment SGi Deployment EC Apps EC Platforms Data Network UPF UP EC Apps EC Platforms 4G S1u Deployment Non-standalone 5G Standalone 5G
- 5. Mobile Edge Platform HTTP Server ME App ME App Service Agent App External API NES API NES Thread (User space Device Driver) NIS MEC Manager Redis Database HTTP Server Process NES Process Redis Process NTS Service Registry Legend Dependence Module Scope of the Mobile Edge Platform Scope of the NEV SDK High Level Design Platform Producer Linux Tap MEC API MEC API MEC API DNS Msq. MEC API MEC API Modules run a root NEV SDK Data Plane NEV SDK APIs NEV SDK components interaction 5 Local Host HiRedis Lib: Socket DPDK Ring HTTPS REST HTTPS REST HTTPS REST Local host HiRedis Lib: Socket VirtIO/vhost/ SW Ring FCGI Socket Web Socket C API call Unix Socket Data traffic NFV Infrastructure Optimizations
- 6. NEV SDK: Data Center Cloud and Cloud Native Opens Stack + EC Manager EC Apps Policy and lifecycle management mm5 App NEV SDK Kubernetes + EC Manager EC Apps Policy and lifecycle management mm5 NTSMicroservicesPod OVS/VPPPod MECProducerAppPod AmazonCloud AdapterPoD AzureCloud AdapterPoD MECConsumerAppPod NES Microservices Pod Mm5 API Mp1 API MECConsumerAppPod MEC Service VM Intel Architecture Advantages: Multicore Architecture, Intel FPGA, Intel Memory, Virtualization enhancements…
- 7. Open Source: Available on Intel github https://github.com/IntelCorporation | NFD at https://github.com/kubernetes-incubator/node-feature-discovery SOFTWARE AVAILABILITY* Challengesbeingaddressed Solution Address key infrastructure challenges in Cloudnative Edge Cloud deployment Ability to request/allocate platform capabilities CPU Core-Pinning and isolation for K8s pods Dynamic Huge Page allocation Platform telemetry information Discovery, Advertise, schedule and manage devices with K8s Guarantee NUMA node resource alignment Fail-over and high availability of networking High performance Data Plane (N-S) High performance Data Plane (E-W) Multiple network interfaces for VNFs Node Feature Discovery CPU Manager for Kubernetes Native Huge page support for Kubernetes BOND-CNI Device Plugin NUMA Manager VHOST USER SR-IOV Open Source: Nov. ’16 Upstream K8: Incubation Graduation TBD Open Source: V1.2 April ’17 Upstream K8: Phase 1 - V1.8 Sept ‘17 Upstream K8: V1.8 Sept ‘17 Open Source: CNI plug-in – V1.0 Dec ‘17 Upstream K8: Phase 1 - V1.8 Alpha Upstream: Working PoC with proposal Upstream K8: TBD Open Source: CNI plug-in - V2.0 June ‘17 Upstream K8s: TBDOpen Source: CNI plug-in – V1.0 Sep ‘17 Open Source: CNI plug-in – V2.0 April ‘17 Upstream collectd: V5.7.2 June ‘17 ; 5.8.0 ((Q4 2017 date TBD) Switchingandlow latencynetworkingReal-timeCompute
- 8. 8 Overview of ME application Instantiation APIs in NEV SDK
- 9. Overview of ME application APIs in NEV SDK
- 10. 10 ScatterRecv NIS_UPSTR_GTPUC NIS_UPSTR_SCTP NIS_UPSTR_RNIS NIS_UPSTR_GTPC Portx/ Rx NTS_UPSTR_IP Porty_IO_ANY EnqDeq ClassifyDeq Enq IO_VMx_ANY Porty/ Tx Send Send VMx EnqDeq EnqDeq EnqDeq Portz_IO_ANY Send Portz LBP Default route NES NTS NIS Legend: Default route NTS_UPSTR_GTPU ClassifyDeq Enq ScatterRecv NIS_DWSTR_GTPUC NIS_DWSTR_SCTP NIS_DWSTR_RNIS NIS_DWSTR_GTPC Portx/ Rx NTS_DWSTR_IP Porty_IO_ANY EnqDeq ClassifyDeq Enq IO_VMx_ANY Porty/ Tx Send Send VMx EnqDeq EnqDeq EnqDeq Portz_IO_ANY Send Portz LBP Default route NES NTS NIS Legend: Default route NTS_DWSTR_GTPU ClassifyDeq Enq NEV SDK Flexible Data plane VM/LBP flows Downstream flowUpstream flow
- 11. eMBB/Mobile devices Base station LTE mMTC IPv4 NBIOT/5G NR Base station AR compute local cloud EPC Internet Or from CUPS IOT G/W and modules NEV SDK Amazon Green Grass Cloud adapter overview MEC GG adapter Amazon AWS IOT G/W IOT Hub Azure Resources *Other names and brand may be claimed as the property of others.
- 12. Intel® NEV Roadmap Network Edge Virtualization SDK Market Segments Industrial, Venues (Stadiums, Cities…), Healthcare, Retail, Vehicles, IOT ONAP Openstack, Kubernates Orchestration Intel Edge Platform Manager Virtual RADIO connectivity (4G/5G NR, wifi) FlexCore (5G) vEPC Core VNFs Intel vCPE VNFs Edge Apps: CDN, AR/VR, IoT Analytics, Connected Cars, Intel Cloud Adapter Akraino – Wind River* Titanium Cloud (Linux*, AVS, OpenStack/Kubernetes/Docker* VIM) Networking Intel® Xeon Based Server FPGA Storage Cloud Apps Intel NEV Reference Edge Platform Example Cloud Portfolio: Functions, Event handlers, Function Stores, Function RTE… R1.0 (2015) Intel Atom® Processor Kit Ease of Use Virtualization Environment Intel® System Studio (ISS) Tool Suite R2.0, 2.1 (2016) Intel® Xeon® Processor Kit ETSI MEC Like RESTFUL API Radio Network Information Service (RNIS) DNS Caching NFV Tool Kit with Control and Compute Node 3rd party UE/eNB/EPC emulators 1H 2018 Multi-Access LTE, WiFi, Wired & Etc GG & Azure CentOS*, Wind River* Titanium S1 Interface SGI LAN Interface 2H 2018 *OpenVINO & nGRAPH Policy, telemetrics, Containers, K8 Enhanced KPIs, CDN, Gaming sample Apps GPU, Movidus, & FPGA Plug Ins *Intel’s nGraph compiler offers “framework abstraction” for the growing wave of developers and data scientists coming into AI, while the OpenVINO SDK for developers looking to add deep learning and computer vision 1H 2019 Cluster enablement, Analytics, AI libs Optimizations OpenAPIs Network Slicing 5G NR deployment support
- 13. 13 Parking management Over the air (ota) update – Intel and CarSync* HD maps Remote control Traffic management Mobile platooning Entertainment & VR Edge cloud 5G Edge cloud 5G Routers / core network Cloud / data center On-Premises Data Center 5G mmwave High volume data offload Cameras/ parking/ presence *Other names and brand may be claimed as the property of others.
- 14. • Average ~30s savings with MEC in internet live broadcast (versus other OTT) • Stadium audience (CUC, Nokia) – enjoy multi angle viewing experience 0.5s latency • General Internet audience (Nokia, Tencent) – 30s latency advantage with another internet live broadcast provider Smart Stadium at Shanghai Mercedes Benz Arena Stadium audience Internet audience Cloud On-premp Local Breakout *Other names and brand may be claimed as the property of others.
- 15. Warehouse&Manufacturing 15 EdgeCloudGaming Inspection&Infrastructure SmartStadiums VR/ARwithlocation Security,Recognition, Surveillance ADAS Retail SmartCity Hospitality&Casinos TRANSPORTATION
- 16. Resources
- 17. Links and Resources • Intel NEV SDK • https://networkbuilders.intel.com/network-technologies/nev • Intel Parallel Studio • https://software.intel.com/en-us/intel-parallel-studio-xe
- 18. Glossary • NFV – Network Function Virtualization • VNF – Virtual Network Function • SDN - Software Defined Network • 3GPP – 3G Partnership Project • RAN – Radio Access Network • vRAN – Virtualized Radio Access Network • FlexRAN – Flexible Radio Access Network • SDK – Software Development Kit • EPC – Evolved Packet Core • MEC – Multi-Access Edge Compute • NEV – Network Edge Virtualization • eNodeB / eNB – Enhanced NodeB (BaseStation) • UE – User Equipment (Device) • COTS – Commercial Off The Shelf • CPRI – Common Public Radio Interface • RRH – Remote Radio Head • FPGA – Field Programmable Gate Array (Accelerator) • PHY – Physical Layer (L1) • MAC – Medium Access Control (L2) • RLC – Radio Link Control (L2) • RRC – Radio Resource Control (L3) • PDCP – Packet Data Control Protocol (L3) • TNL – Transport Network Layer • RF – Radio Front End • SON – Self Organizing Network • OAM - Operation Administration And Maintenance • QAT – Quick Assist Technology • DPDK – Data Plane Development Kit • DL – DownLink (eNodeB to UE) • UL – UpLink (UE to eNodeB) • OAMP – Operations, Administration, Maintenance and Provisioning • NIC – Network Interface Card • VM – Virtual Machine • KVM – Kernel Virtual Machine • 5G TF – 5G Technical Forum • 5G NR – 5G New Radio • IA – Intel® Architecture • NDA – Non Disclosure Agreement • ETSI - The European Telecommunications Standards Institute
- 19. Disclaimer Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance varies depending on system configuration. No computer system can be absolutely secure. Check with your system manufacturer or retailer or learn more at www.intel.com. Cost reduction scenarios described are intended as examples of how a given Intel- based product, in the specified circumstances and configurations, may affect future costs and provide cost savings. Circumstances will vary. Intel does not guarantee any costs or cost reduction. Intel, the Intel logo, are trademarks of Intel Corporation or its subsidiaries in the U.S. and/or other countries. *Other names and brands may be claimed as the property of others. © Intel Corporation