Dr.-Ing. Leonardo Gomes Baltar () Senior Standards and Research Engineer with inputs from Dario Sabella, Soo Jin Tan and ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Dr.-Ing. Leonardo Gomes Baltar (leo.baltar@intel.com) Senior Standards and Research Engineer with inputs from Dario Sabella, Soo Jin Tan and Prakash Kartha
At Intel We’re powering the future of computing and communications, delivering experiences once thought to be impossible. Autonomous 5g Artificial Virtual Driving networks intelligence Worlds 2
Intel technology leadership Make the world’s best semiconductors Be the Leading end to end platform provider for the new data world Lead the Artificial intelligence & Autonomous revolution Deliver the best customer experiences on the planet 3
Semiconductor leadership enables Platform leadership CPU FPGA SoC Co-optimized Co-optimized process and products product design MEMORY MODEM Intel’s Integrated Device Manufacturer model is foundational to delivering best-in-class platforms 4
Intel delivers end to end Platforms Devices/edge NETWORK Cloud/Data center CUSTOM MemOry 4g/5G Software 5
powering AI computing from edge to cloud Combining Intel hardware and software to accelerate AI workloads for data-intensive tasks like autonomous driving Devices / edge network Cloud/data center 6
Contents • Motivation • What is Edge Computing? • Examples of Edge Computing in road ITS • In-Vehicle and City Experiences– MWC 2018 • Smart RSU example – MWC 2019 • Edge Computing standardization and ecosystem • ETSI ISG MEC • 3GPP • 5GAA • AECC • Challenges 7
Transformation of Automotive Industry … Transformation of Automotive Industry in Upcoming Decades Advances in electronic and computing opened new transportation opportunities Automotive sector transforms to deliver safe, green and intelligent transportation with ultimate goal to enable self-driving cars which are becoming reality Evolution of Human Transportation Reality Outlook or Fantasy History Ongoing Transformation 9
…Toward a data driven era Practical Challenges Additional Component Human sensing / control capability • Cellular communication Environment perception through sensors: cameras, radar, LIDAR, and centralized Future automated driving is to be further processing of sensor data to control vehicle empowered by cellular communication behavior technologies Real time massive data processing Parallel / sequential processing Core Technology Components Artificial intelligence Sense / fuse / decide High-end computing capabilities Latency / big data / throughput (computing and memory advances) Precise location / navigation 10
Automated and Connected cars – key drivers Connected Car Vision Cloud V2X services Over the air updates Infotainment / media delivery Intelligent route and path planning Tracking / fleet management Transportation as a service Inter-Car Communication Vehicles talk to another cars, pedestrians, road side units Road safety Telematics information exchange Environment perception Cooperative & automated driving 11
12
WHAT IS EDGE COMPUTING? EDGE COMPUTING IS THE PLACEMENT OF ENDPOINT DATA CENTER-GRADE NETWORK, COMPUTE & STORAGE Closer to DEVICES TO IMPROVE SERVICE Optimize Comply with Data And reduce application CAPABILITIES TCO Locality Latency Outmost BEFORE TRANSITION layers of The Edge TO THE ENDPOINT Is the Processing OR Another Network or Network 13
DATA WORKLOADS DRIVE EDGE COMPUTING VIDEO WIRELINE HEALTHCARE FIXED ACCESS (vCCAP, PON) SMALL CELLS MANUFACTURING LICENSED & DRONES EDGE ENERGY UNLICENSED Driver for edge DATA CENTERS RETAIL Latency, bandwidth, Security, connectivity NEXT GEN TRANSPORTATION CENTRAL OFFICE (vEPC, vBNG) CUSTOMER PREMISES SMART EQUIPMENT PCs CITIES (uCPE, SD-WAN, WIRELESS ACCESS EDGE COMPUTE BASESTATIONS & OR FOG NODES) RAN PHONES Devices / On-premises Network Network Data Center| Things edge edge Core Cloud Latency expectation Varies
Market Forecasts • MEC incorporates the benefits of virtualization and cloud-computing in order to place high-powered computing capabilities as close as possible to subscribers. • The largest number of installations, according to iGR (2017), will be for retail and health care and social assistance. And this forecast was still not considering other emerging verticals (e.g. VR/AR, automotive, industrial automation, …) Distributed Cloud unlocks 25% of the 5G operator revenue potential (*) (*) Source: Ericsson study on 5G business potential 15
MEC KPIs and their impact on business 11. Latency improvement 22. Network utilization and cost savings 1 2 33. Energy efficiency, TCO analysis 44. Computational and Networking resources • Motivation for these KPIs comes from 4 e.g. current positioning of operators (*) 2 3 • Also importantly, performance metrics are applications-specific as they are tailored to vertical segments (e.g., automotive, industrial automation, VR/ AR etc.). • Performance assessment of MEC is key. (*) A. Manzalini , “Multi-access Edge Computing: Decoupling IaaS-PaaS for enabling new global ecosystems”, Berlin 19-20 Sept 2018, Edge Computing Congress, 2018 16
Where is the edge? MEC performance should be compared against different deployment options MEC gains depends also on the vertical / use case app app Connected macro vehicles app app app Web RRH RRH Remote Cloud Industrial RRH BBU automation app app Global Level E-Health micro devices micro National Level Local/Regional Level Access Level VR/AR devices > 10000x 100x 1000x 10x 100x 1x 10x Scale 17
18
Advanced V2X & Edge Computing – MWC 2019 DEMO 19
Architecture Overview – MWC 2019 DEMO Next Gen MEC Controller Sidelink OTA Next Gen V2X on V2X on Intel Intel® Xeon® D Cloud Atom™ IVI Processor Platform System V2X Others Analytics APP Car Unit Smart RSU Cloud Next Gen V2X Intel® Open NESS OpenVINO™ (coming soon) (optional) toolkit Intel® Movidius™ CPU Storage Networking Myriad™ X VPU (optional) Smart RSU PoC SW Architecture High Level PoC Architecture More on Open NESS at www.open-ness.org
Next Gen V2X Wireless Test Platform – MWC 2019 Demo Key features V2V, V2P, V2I radio links based on 5G NR technology Test chip with integrated communication modem and neural network capabilities MEC and FlexRAN for network infrastructure Neighborhood vehicles and environment as compute/storage nodes to form dynamic local cloud Enable new wave of time sensitive and local relevance service in addition to V2X usages
22
ETSI ISG Multi-Access Edge Computing (MEC) • MEC technology - previously Mobile-Edge Computing – offers IT service and Cloud computing capabilities at the edge of the network • Develops specifications that will allow the hosting of third-party applications in a multi- vendor Multi-Access Edge Computing (MEC) environment and accelerating the development of edge applications across the industry, to increase market scale and improve market economics • By shifting processing power away from remote data centers and closer to the end user, it enables an environment that is characterized by proximity and ultra-low latency, and provides exposure to real-time network and context information ETSI MEC phase 1 ETSI MEC phase 2 End 2014 End 2016 23
ETSI ISG Multi-Access Edge Computing (MEC) • Phase 1 - Platform Service API specifications, such as Application Enablement, Radio Network Information Service, Location Service, User Equipment Identity Service, Bandwidth Management Service and User Equipment Application Enablement • Phase 2 - New use cases and requirements to address topics such as charging, regulatory compliance, support of mobility, support of containers, support of non-3GPP mobile and fixed networks, slicing, 3GPP 5G MEC integration, testing, and key vertical use cases such as automotive. Integration with ETSI’s ISG on Network Functions Virtualisation (NFV) ETSI MEC phase 1 ETSI MEC phase 2 End 2014 End 2016 24
ETSI MEC standard work on V2X • MEC GR-022 - support for V2X use cases Focuses on identifying the necessary MEC support for V2X applications Collects and analyze the relevant V2X use cases (e.g., the from external organizations.) And evaluate the gaps from the defined MEC features and functions, by identifying the new requirements including new features and functions • MEC GS-030 - V2X service API Introduces new MEC Service, V2X Information Service (VIS) based on the study report conclusions in MEC GR 022 Support for service exposure and communication ability across participating V2X applications in different MEC systems Expose PC5 interface configuration parameters of one system to another system (the right-hand side figure is showing a relevant multi-operator scenario) 25
Edge computing in 3GPP • 3GPP has adopted the Edge Computing concept as part of their 5G architecture • 3GPP TS 23.501 (Clause 5.13) on the architecture for 5G Systems, contains a set of new functional enablers for the integration of MEC in 5G networks. • 3GPP is also developing the API framework for the northbound APIs, taking into account corresponding ETSI ISG MEC APIs • ETSI White Paper No. 28 “MEC in 5G networks” • Focuses primarily on exploring edge computing enablers from 3GPP and on how the 3GPP ecosystem may benefit from the MEC system and its APIs, as a set of complementary capabilities to enable applications and services environments at the edge of mobile networks 26
5G Automotive Association – 5GAA 27
5G Automotive Association 28
5G Automotive Association • Edge Computing as priority area • White paper: “Toward fully connected vehicles: Edge computing for advanced automotive communications” available at www.5gaa.org • Covering use cases, business models, standardization and ecosystem • Cooperation agreement with ETSI • Particular interest in collaboration with ETSI ITS and ETSI MEC • More to come in the near future.... • New Work Items on key areas such as MEC and VRU 29
Automotive Edge Computing Consortium - AECC Network/Computing for Global and Initial Focus Leading Market Actors Automotive BIG DATA Sustainable Ecosystem on Vehicle to Cloud Join Forces To accommodate the The next generation The context is the high Toyota, Toyota ITC, Ericsson, automotive big data in a connected cars will require volume Vehicle to Cloud Intel, Denso, DOCOMO, KDDI, smart and efficient fashion. improved systems in order to services and not the classical NTT are funding members evolve. V2X/C-ITS services with an and more already joined or initial focus on automated are in the pipe to join vehicles Automotive service evolution needs BIG DATA and its infrastructure 30
AECC Vision and Challenges SAFETY ITS Evolution of • P2P (sidelink) • Network architectures • Low latency • Computing infrastructures • High reachability • Business models Short Range CELLULAR V2X USER EXPERIENCE IVI BIG DATA CAPACITY • Cloud (downlink) Accommodate automotive • Cloud (uplink) • High Interactivity • Big data capacity CHALLENGE! data & services • Delay tolerant • Smart and effective fashion • Multi-operator environment • Multi-edge environment CELLULAR WiFi CELLULAR Challenge to get a homogenous service in a heterogeneous network 31
AECC – System-wide Service Requirements Source: AECC White Paper available at www.aecc.org 32
33
Edge computing Challenges for Road ITS • Support and interoperability • Multiple road and network infrastructures • E.g. Multiple RSU and cellular network operators • Multiple infrastructure vendors • Multiple applications and service providers • Multiple vehicle OEMs and suppliers • Support global interoperability • Effective incorporatio of AI enabling technologies • Appropriate Radio Access Neworks • Functional Safety Compliance 34
“Don’t be encumbered by history, go off and do something wonderful.” Intel cofounder Bob Noyce 35
© Intel Corporation *Other names and brands may be claimed as the property of others. | Intel, the Intel logo, and XMM™ are trademarks of Intel Corporation in the U.S. and/or other countries. Intel, the Intel logo, and XMM™ are trademarks of Intel Corporation in the U.S. and/or other countries. | 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.
You can also read