WIXLIB

Part 2: Predictive QoS for Edge Computing: Insights from 5GAAPresentation at the20th March 2020Dario Sabella(INTEL, Next Generation and Standards, Vice-Chairman of MEC)Web & Networks IG

AgendaIntro on edge computing (https://www.w3.org/wiki/Networks)MEC V2X API and Predictive QoS Use cases of interest 5GAA Predictive QoS MEC demoConclusionsAbstractPart )2Already presented in a past W3C conf callSlides here: 020Jan/att-0001/01-partETSI MEC standard Part 1Munich, 15-01-2020 – Presentation to W3C – Web & Networks IGThe talk will be focused on Intel activities on edgecomputing, ranging from standardization (e.g. ETSI MEC),industry groups (e.g. 5GAA – 5G Automotive Association)and mentioning also some PoCs and experimental activities.Particular focus will be given to MEC use cases of potentialinterest for W3C (e.g. application offloading), and on Intelengagement in the automotive domain (mentioning a recent5GAA Demo on MEC and predictive QoS).

5GAA(5G Automotive Association)

Driving Connected Mobility Forward5GAA brings together automotive, technology and telecommunications companiesto work closely together to develop end-to-end connectivity solutionsfor future mobility and transportation servicesAUTOMOTIVE INDUSTRYTELECOMMUNICATIONSVehicle Platform, Hardwareand Software SolutionsConnectivity and NetworkingSystems, Devices and Technologies5GAA unites today over 130 members from around the world working together onall aspects of C-V2X including technology, standards, spectrum, policy, regulations,testing, business models and go-to-marketMunich, 15-01-2020 – Presentation to W3C – Web & Networks IG4

5GAA Organisational StructureBOARDGENERAL ASSEMBLYEXECUTIVE COMMITTEEWG1WG2WG3WG4WG5WG6WG7Use Cases andTechnicalRequirementsSystemArchitecture andSolutionDevelopmentEvaluation,Testbeds and PilotsStandards andSpectrumBusiness Modelsand Go-To-MarketStrategiesRegulatory andPublic AffairsSecurity andPrivacyMunich, 15-01-2020 – Presentation to W3C – Web & Networks IG

What is C-V2X (Cellular-Vehicle to Everything)?A comprehensive road safety and traffic efficiencysolution that allows vehicles to communicate with:Mobile networks (V2N)Road Infrastructure (V2I)Cyclists via smartphones (V2P)Pedestrians viasmartphones (V2P)Other vehicles (V2V)Munich, 15-01-2020 – Presentation to W3C – Web & Networks IG

Supporting innovation:Leverage solutions together On-going dialogue between 5GAA and leadingtechnology partner organisations to advanceinnovation in all world regions 5GAA to continue strengthen technicalcooperation and maximise benefits with existingorganisations or projects by reinforcingengagement via common members Develop new partnerships with relevantorganisations on specific aspects e.g. vulnerableroad users or road authorities7

Automated and Connected cars – key driversConnected Car Vision Cloud V2X servicesOver the air updatesInfotainment / media deliveryIntelligent route and path planningTracking / fleet managementTransportation as a serviceMEC is a key technologyfor many of these driversInter-Car Communication Cars talk to another cars, pedestrians, road sideunits Road safety Telematics information exchange Environment perception Cooperative & automated drivingSuggested reading: 5GAA White Paper on Edge 2/5GAA T-170219-whitepaper-EdgeComputing 5GAA.pdfMunich, 15-01-2020 – Presentation to W3C – Web & Networks IG8

Predictive QoS in 5GAA 5G Automotive Association (5GAA) has developed the concept of predictive Quality of Service (QoS), which is a mechanismenabling mobile networks to provide advance notifications about predicted QoS changes to interested consumers. This makes it possible to adjust application behaviour before the predicted QoS change takes effect, which is important tocertain automotive use cases, such as remote and autonomous driving. QoS predictions may also be exposed to the V2X Application through a MEC service APIUse CaseTele-Operated DrivingQoS KPIs to be PredictedData rate, Latency, ReliabilityHigh-Density PlatooningLatency, ReliabilityHazardous Location WarningReliabilityLane MergeLatency, ReliabilityExamples of Potential Application ReactionsChange route, Park vehicle, Handover to nearbydriver, Change sensor set/properties, Changeteleoperation mode (e.g., from manoeuvring totrajectory provision)Change inter-vehicle distance, Handover to driver,Change platoon speed or length, Terminate platoonInform user about availability of warning service,Change speed, Change routeChange speed of merging attempt, Abort lane mergeSoftware UpdateInfotainmentData rateData rateReschedule, Stop or resume downloadChange video quality Work Item MEC 030 introducesnew MEC Service, called V2XInformation Service (VIS). Key requirement for MEC V2X APIis the support for service exposureand communication ability acrossparticipating V2X applications indifferent MEC systems Stable draft available at the MECOpen Area:https://docbox.etsi.org/ISG/MEC/OpenSuggested reading: 5GAA White Paper on “Making 5G Proactive and Predictive for the Automotive 0/01/5GAA White-Paper Proactive-and-Predictive v04 8-Jan.-2020-003.pdf9Munich, 15-01-2020 – Presentation to W3C – Web & Networks IG

5GAAIntel Demo on x-as-a-market-reality/

In-Vehicle Entertainmentutilizing Multi-AccessEdge Computing (MEC)over 5G Networks Multi-Access Edge Computing (MEC)is a key technology offering cloud computingcapabilities and an IT service environment atthe edge of 5G mobile networks. This demo, realized by Intel, together with Marelli,Terranet and Equinix, shows how MEC can supportimmersive high-definition (HD) entertainment for alloccupants of a moving vehicle, including videostreaming, gaming, virtual reality (VR), office work,online education, advertisement, etc.[figure: Continental AG]11

In-Vehicle Entertainment – motivation in 5GAAAutomotive sector is key for MEC and for 5G.These technologies will open the door to multiple use cases andservices that can be monetized by 5GAA stakeholders.In addition to traditional use cases on connected andautomated cars, infotainment is also a promising market.(Excerpt from Intel 5G Connected Vehicles Webinar):"Intel predicts a new 7 trillion passenger economy will emergewhen passengers become riders." [ ]"We did a study that shows drivers spend 300 hours a year behindthe wheel and 5G offers entertainment opportunities to optimizethat time as we transition from drivers to riders."[figure: Intel]12

Predictive QoS – reference scenariovehicle planning to follow a trajectoryfrom location A to location B ABExample of map @ Device/car DisplayMunich, 15-01-2020 – Presentation to W3C – Web & Networks IG

Predictive QoS – reference scenariopredicted QoS info examples:1 - Resource Type (GBR, delaycritical GBR or Non-GBR);2 - Priority level;3 - Packet Delay Budget;4 - Packet Error Rate.5 - Averaging window.vehicle planning to follow a trajectoryfrom location A to location BMEC could be used to support infodelivery of radio conditions “en route”,ahead of the vehicle’s passing time,before reaching a decision. ITSclientappAP (x,y,t)Edge CloudRAN/RSUITSMECappMunich, 15-01-2020 – Presentation to W3C – Web & Networks IGB

Predictive QoS – reference scenariovehicle planning to follow a trajectoryfrom location A to location BMEC could be used to support infodelivery of radio conditions “en route”,ahead of the vehicle’s passing time,before reaching a decision.Decisions may consist in e.g.,downloading infotainment content,enabling/disabling autonomous drivingfeatures, scheduling SW updates, etc. ITSclientappAP (x,y,t)Edge CloudRAN/RSUMunich, 15-01-2020 – Presentation to W3C – Web & Networks IGVideostream/ contentsB

Predictive QoS – reference scenariovehicle planning to follow a trajectoryfrom location A to location BMEC could be used to support infodelivery of radio conditions “en route”,ahead of the vehicle’s passing time,before reaching a decision.Decisions may consist in e.g.,downloading infotainment content,enabling/disabling autonomous drivingfeatures, scheduling SW updates, etc. ABUse Case 1 “coverage”Use Case 2 “coverage”Use case 3 “coverage”Munich, 15-01-2020 – Presentation to W3C – Web & Networks IG

In-Vehicle Entertainment – demosThis demo (based on IVE use case) aims at demonstrating two main concepts:MECApp The benefits of edge computing, bycomparing different deployment optionsRemoteAppUEAppwwwRemote AppserverRAN Edge CloudMECAppRemoteAppUEAppwwwVideo stream/ contentsRAN Edge Cloud[Realized by Marelli/Intel/Equinix] The multimedia delivery adaptation,thanks to Predictive QoS through realtimesimulation and live data from drive testsClient AppRemote AppserverMEC ServerSimuLTE/5G NetworkPGWUugNBXMedia Server[Realized by Terranet/Intel/UniPisa]17

In-Vehicle Entertainment – demosThis demo (based on IVE use case) aims at demonstrating two main concepts:MECApp The benefits of edge computing, bycomparing different deployment optionsUEAppRemote AppserverRAN Edge CloudMECAppUEApp[Realized by o stream/ contentsRAN Edge CloudRemote Appserver The multimedia delivery adaptation,thanks to Predictive QoS through realtimesimulation and live data from drive tests[Realized by Terranet/Intel/UniPisa]18

Benefits of Edge DeploymentsMECAppUEAppRemoteAppwwwRemote AppserverRAN Edge CloudMECAppUEAppRwwwVideo stream/ contentsRAN Edge CloudRemote Aserver The IVE client App running on the On-Board Unit (OBU) displays high-definition video contents to passengers basedboth on their interests and car context (e.g. position, heading, speed, etc.). Video contents are retrieved from off-board services in the Cloud or at the Edge. The demo shows how a media server instantiated at the edge can provide lower communication latencies, fasterservice response times, higher bandwidth and more dynamic and fine-grained geographical provisioning.19

Reference Architecture used for IVE App TestsOn-boardIVE MECOff-board ServiceCellularConnectivityOn-Board Unit (OBU)Edge Node (MEC) Runtime(*)IVE MECOn-Board AppMEC Edge Node (off-board)Frankfurt NodeDemoUser InterfaceIVE MECOff-board ServiceTablet or Digital CockpitPublic CloudAmsterdam NodeBenchmark NodesIVE MECOff-board ServicePublic CloudUSA Node(*) the usage of OpenNESS in the 5G live network is foreseen for a next phase of the trial activities21

MEC performance evaluation – preliminary comparative resultsDrive tests in Turin with 3 serverlocations: FrankfurtAmsterdamUSA23

In-Vehicle Entertainment – demosThis demo (based on IVE use case) aims at demonstrating two main concepts: The benefits of edge computing, bycomparing different deployment options[Realized by Marelli/Intel/Equinix] The multimedia delivery adaptation,thanks to Predictive QoS through realtimesimulation and live data from drive testsClient AppMEC ServerSimuLTE/5G NetworkPGWUugNBXMedia Server[Realized by Terranet/Intel/UniPisa]24

Collecting Data from Live Measurements in Turin Cellular network quality LOG file wascaptured during drive test in Turinwith Terranet softwareLOGfile Terranet provided client application,media server backend and cellularnetwork visualizationSimuLTE/5G NetworkPGWXUu Realtime emulation recreates networkquality from actual test drivegNB25

Test Drive Routes in TurinPath ofthe demoTurinSmart RoadcircuitSmart Road circuit26