Connected Cars: on the Edge of a Breakthrough How the AECC’S Distributed Edge Architecture Can Drive Revenue for Mnos

Home , Connected car, Mobile World

WHITEPAPER Connected Cars: On The Edge Of A Breakthrough How the AECC’s distributed edge architecture can drive revenue for MNOs

Published by Executive Summary

Mobile networks will enable many of this decade’s most promising technologies, and perhaps no other device holds as much potential as the connected vehicle. GSMA has forecast that the global connected car market will be worth $198 billion by 2025.1 According to the analyst team at McKinsey, the average vehicle will generate an estimated $310 a year in connectivity revenue by the end of this decade2 (annual smartphone revenue per user (ARPU) would have to grow at a compound annual growth rate (CAGR) of more than 10% to match that3).

As each vehicle on the road becomes part of at least one mobile network, the opportunity for operators is significant. So, too, is the cost of letting the connected car ecosystem evolve without coordinated cooperation from mobile network operators (MNOs). By 2025, at least 400 million passenger vehicles globally will have embedded connectivity, according to SBD Automotive's 2020 Connected Services Forecast.4 A data deluge from connected cars is coming no matter what, and operators that get involved now will be ready to handle the data more economically and efficiently.

The AECC is a multiindustry collaboration formed to maximize the opportunities created by vehicular connectivity and automotive big data. Recognizing the connected car’s reliance on cellular networks, leaders from the automotive industry are partnering with carriers, equipment vendors, software developers and cloud service providers.

Proactive mobile network operators are collaborating with the auto industry, which is increasingly focused on delivering services in addition to simply selling vehicles. Services such as dynamic HD maps and intelligent driving rely on the network’s ability to instantaneously process localized data and transmit actionable insights. Mobile networks will need the computing resources to process the right data in the right place at the right time: a topology aware distributed edge computing cloud architecture.

This white paper examines the opportunities for MNOs to become leaders in the emerging ecosystem of the automotive edge. To fully capitalize on this opportunity, mobile network operators need to work closely with a consortium of automotive, edge and cloud solution providers.

2 WHITEPAPER The AECC’s Topology-aware Distributed Edge Computing Cloud Architecture

equipped with fast internet access, With the right infrastructure and What is the artificial intelligence and access to protocols in place, connected cars AECC and how is its big data analytics for high-definition will be able to interact with cellular map creation and distribution, as networks rather than just using them approach unique? well as for services like intelligent to send massive amounts of data to driving and more. Within four years, the cloud. Navigation is a good The AECC and its members are hundreds of millions of vehicles will example. Unlike today’s navigation working to identify the connected be sending a thousand times more apps, which use relatively small vehicle requirements needed to data to the cloud than today’s amounts of cellular data, real-time design the next generation of connected cars. high-definition maps will be in networks that can think globally and constant communication with the act locally. As cars become As the connected car market rapidly network, providing a service that will increasingly autonomous, data will expands beyond luxury models and be highly valuable to drivers. Cars be processed by the vehicles, in the premium brands to high-volume, will be able to quickly alert drivers to cloud and in between at mid-market models, the industry dangerous traffic patterns or help strategically located edge compute will soon reach a tipping point. The them find parking in crowded areas. locations that will capture data that volume of vehicle data generated does not need to be aggregated in will overwhelm existing cloud, A localized distributed edge the cloud. computing and communications computing cloud architecture infrastructure resources. speeds data processing by matching Consider the numbers: with a billion compute resources with the cars already on the planet, each Not all vehicular data needs to vehicles that need them. In the vehicle replacement represents a travel to the cloud. “Data has a context of edge computing for new connection. By 2025, 100% of nature of locality,” explained automotive use, the “edge” means all new vehicles, more than 100 Toyota’s Ken-ichi Murata, Chair of the hierarchically distributed non- million, will come to market the AECC, and Fellow and General central clouds where computation connected, according to IHS, Manager of Connected Cars at resources are deployed, and edge bringing new services and business Toyota. “The vehicles in Texas do not computing technology can be used models to bear. This next generation need data from New York. … That’s to design a flexible topology-aware of connected cars will need to be just a waste of network.” cloud infrastructure.

Connected Cars: On The Edge Of A Breakthrough 3 Distributed edge computing on localized networks

Cloud Data Distributed Computing Data #-$.) Data Local Data Integration Platform Reduces concentration of computation to shorten Inter-connects local cars to processing time Local Local integrate information Data Data Local Local Data Local Local Data Data Data

Intelligent High Intelligent High V2C2V Driving Definition V2C2V Driving Definition !" Map !" Map Distributed Computing Distributed Computing Localized Network Localized Network Localized Network Splits data traffic in order to handle huge data flow

Connected Car Connected Car Connected Car Connected Car (Source: AECC)

This architecture creates an AECC Director Roger Berg, Vice thousands of companies that service opportunity for mobile network President for R&D in North cars and trucks, all of which can operators to host edge servers America at DENSO International benefit from automotive big data. running cloud-based applications. America, sees smartphones already Makers of roadside infrastructure are The cloud is dominated by a handful creating a need for a robust edge also part of the mobility ecosystem, of hyperscalers, but at the edge, network. “If similar needs from the since street signs, billboards and MNOs have more of a chance to mobility industry can be met by a stoplights can all potentially be participate, given their position as distributed computing network connected to the internet. owners of the network access architecture set up for smartphone points. MNOs are already apps, then the MNO has another Many different technology experimenting with this business vertical from which to gather providers are investing human and model, according to the AECC’s additional ARPU sources from the financial capital to learn how to Proof of Concept Committee. same network,” Berg noted. better serve the automotive industry. From the chips that As mobile operators invest in edge- connect devices and power servers based servers meant to host cloud The Distributed to the data centers that process applications, the goal is typically to Edge Ecosystem exabytes per day, IT hardware is support mobile subscribers. But integral to the connected car. In within a few short years, this These futuristic features will rely on addition to mobile network infrastructure could support not the cooperation of many different operators, fiber providers, satellite just mobile but also mobility, the ecosystem participants. Connected operators and Wi-Fi equipment connected car ecosystem that auto vehicles are supported not just by the vendors are eyeing the connected executives call the mobility industry. automakers, but also by the car opportunity.

4 WHITEPAPER Connected vehicles are supported by a heterogeneous ecosystem

Wide-area Roadside cellular network equipment

Fleet OEM

Wi-Fi

TTelematicse Passenger Wi-Fi Telematics, ADAS, C-ITS

Commercial vehicle for people transport Private vehicle

(Source: AECC and AECC member Ericsson, 2021)

The AECC has assembled a and Toyota, along with information Google. The consortium also partnership that connects and communications technology includes software giants Oracle and industries and brings together the (ICT) leaders AT&T, Dell EMC, Microsoft, which has recently made connected vehicle ecosystem. Cisco, Ericsson, Intel, KDDI, NTT two significant acquisitions of AECC members include automotive and Samsung. Data center leader companies that develop mobile technology heavyweights DENSO Equinix is also a member, as is network technology.

Exploring the Market Opportunity

Some companies will participate in and repurposed, and it will Network operators may have a role to the market solely by selling generate significant value for a play in monetizing automotive data, products or services, while others number of industries. especially if they host edge cloud will want to monetize the data servers to help process this data. But generated by connected cars. “I expect multiple sources of data … for operators, hosting applications at Ownership of that data is a will be used to build a multi-use the network edge is likely to be a complex issue, especially since case, multi-sector analytics platform more immediate opportunity. The drivers may naturally assume that for mobility data collected through AECC has identified five primary use some of the data their cars vehicles and related transportation cases that will be enabled by a generate belongs to them. But infrastructure,” said AECC Director topology-aware distributed edge Roger Berg of DENSO. mostSource: of HPE that data will be analyzed computing cloud architecture.

Connected Cars: On The Edge Of A Breakthrough 5 Use Cases

HD maps are dynamic because they This data session portability is a

HD Mapping are constantly updated with cornerstone of several connected An HD map is a frequently updated information collected by the vehicle use cases, and it relies on the map overlaid with various vehicles. This data can be analyzed readiness of mobile network information such as traffic by the vehicles or by an edge server, operators to use uniform and situations, access ways and road but the maps should be updated by transparent protocols. obstructions, with high precision at the servers because they rely on the centimeter level. In the AECC information from multiple vehicles. Vodafone, Ericsson and Toyota are model, dynamic information is To reliably send and receive HD testing a proof of concept of 5G classified into four layers according map data in transit, vehicles need to core session and service continuity, to the time intervals at which be able to start a data session in the also called “make-before-break.” dynamic information changes. vicinity of one edge server and The team found no loss of continue it in another. connectivity with 5G core make-

6 WHITEPAPER Layers of an HD map

Highly Dynamic Layer

UNEXPECTED Transient STRONG RAIN Dynamic OBSTACLE Layer

Transient Static ROAD WORK ACCIDENT FOG Layer LANE CLOSED

Permanent Static Layer

(Source: AECC) before-break gateway switching in together, their cooperation can generated by both the vehicle and its published results and takeaways jump-start the connected car the driver (biometric data). One from joint trials at the Aldenhoven market by giving automakers goal is to give drivers a complete Testing Center.5 confidence in the networks. and dynamic picture of their driving profiles so that exhausted drivers in Technically, session continuity is “If there are differences in unsafe cars get off the road. Edge likely to be achievable, but networks, how can I build one car at servers can instruct the vehicle on operationally it will rely on carrier production time and deliver it to any what data to send to the cloud and cooperation, like many other of these countries, any of these help to pre-process the data before standard network features. “That’s operators?” Boberg said. “I think it is sent. The driver sees near real- a big discussion with our MNO that is the main thing that makes it time performance metrics. friends,” said AECC Director easy … to produce a car, knowing Christer Boberg, director of IoT and the technology I have incorporated A driver monitoring his or her Cloud Technology Strategies at into the car can be used in any performance/alertness on a road Ericsson. “How do we build this operator network I connect to.” trip is likely to move from one edge network of edges to cater to server’s region to the next during consistent behavior on a global Intelligent Driving the journey. The ability to transfer a scale? Maybe not the same provider data session from one edge server or supplier, but it should have the A second key use case enabled by a to another again becomes critical. same methods.” Boberg said that if distributed edge is intelligent operators are willing to work driving, which relies on sensor data

Connected Cars: On The Edge Of A Breakthrough 7 Vehicle groups are highly dynamic,

Insurance V2Cloud with individual vehicles moving The same data that drivers use to Cruise Assist seamlessly from one group to monitor their own performance can another. Group sizes are also be used over time by insurance V2Cloud Cruise Assist is an evolution expected to be fungible, as the set of companies to reward good drivers of intelligent driving into a more vehicles that needs a given piece of with better rates. Data from flexible service model. Here the information will vary with the connected cars is highly valuable to network mediates vehicle-to-vehicle nature of the information. insurers, and the most valuable data communications by integrating can be collected at the scene of an information obtained from AECC Director Said Tabet, Chief accident. This information can be neighboring cars. This mechanism is Architect for Intelligent Connected analyzed in the cloud, but if pre- called the vehicle-to-cloud-to-vehicle Vehicles at Dell Technologies, said processing occurs at the edge, cars service or simply V2C2V. This service that edge computing will be may be able to make prescriptive scenario is especially effective when essential to managing and suggestions to drivers while they used to broadcast information to processing this data in a timely are still at the scene. For example, vehicles that need the same manner. He said that even drivers can be prompted to information, by utilizing the hyperscale cloud providers will photograph skid marks on the road combination of neighboring vehicles, probably end up directing safety- or to document aspects of a collision roadside units and others. related vehicular data to the edge. before the cars are moved.

Connected vehicle services enabled by a distributed, coordinated edge cloud

Integration & Storage Automotive Cloud • Wideareaintegration& Data Center analysis AI Enabler • Permanent storage Edge Computing Enabling AI based applications in Distributed platform distributed computing platform. Distributed Processing for low latency and • Analysis of local data fast computation. Intelligent with wide area High Driving information V2C2V Definition Flexible topology-aware • Real-time data Map distributed clouds processing • Low-latency response Data traffic to connected cars • Local area information Access NW collection, ﬁltering, Camera aggregation, cleansing Data Traffic Data Traffic Data Traffic

Radio Access Technology Local Devices Sensor Roadside Unit wireless connectivity more • Devices for sensing flexible in quality and cost. road condition, detecting objects e.g. pedestrians

Drone Connected Car Connected Car Connected Car Pedestrians

(Source: AECC)

8 WHITEPAPER Another aspect of V2Cloud Cruise

Assist is simply expediting the Mobility-as-a-Service process of moving information from Ridesharing is just one aspect of the vehicle to the cloud. According mobility-as-a-service. Automakers to Tabet, test fleets of autonomous are starting to see cars as vehicles are generating more data generators of ongoing service than today’s network architectures revenue, and in some cases the car can handle – they download it and itself becomes the service. “Within ship cartridges by FedEx to cloud one day you will see a vehicle being providers. This is not a sustainable used as a mobile library or a mobile model; network architectures need store or a clinic,” said Tabet. to evolve. “Mobility as a service is about integrating into the environment.”

Now Is the Time To Join the AECC

Wireless carriers created a the network, the mobile equipment network operators have an multibillion-dollar smartphone could choose the network with the immediate opportunity to be part of industry by standardizing network necessary capabilities and roam to this process, instead of waiting until protocols, and supporting the next that network even if the incumbent the distributed edge architecture wave of innovation will require even MNO/MSP has a network overlay. … has developed and trying to adapt at greater levels of cooperation. MNOs need to assess the value of a later time. Connected vehicles will need to the capital expenditure necessary to move seamlessly from one carrier’s invest in multi-access edge Connected vehicles are about more edge servers to the next, without computing, and stay competitive in than new revenue models for service interruption or loss of data. markets where more than one automakers and service providers. MNO provides service.” Vehicle emissions are one of the “It doesn’t work for automakers and biggest contributors to climate their service providers if different Investments in edge computing change, and traffic accidents claim a networks have different capabilities resources should be made with as life every 25 seconds. These in Dallas than they do in San much information as possible about problems can be mitigated when Antonio,” said AECC Director Roger how those resources will be used. vehicles have a shared perception of Berg of DENSO, adding that mobile The Automotive Edge Computing their environment. The applications service providers who ignore the Consortium will define that will enable this are predicated reasons to work together run a requirements and develop use cases on wireless network significant risk. for emerging mobile devices, with a interoperability and performance. particular focus on the automotive “In the age of eSIM, a mobile user industry, bringing these The wireless industry’s history has equipment provider and requirements and use cases into repeatedly demonstrated the power applications aggregator can virtually standards bodies, industry of collaborative standard-setting instantaneously change MNO/MSP consortia and solution providers. and technology testing. This will network services in a mobile The consortium will also encourage prove true once again as mobile environment,” Berg said. “So for a the development of best practices network operators approach what is given application with certain key for the distributed and layered likely to be the biggest opportunity performance requirements from computing approach. Mobile since the smartphone.

Connected Cars: On The Edge Of A Breakthrough 9 About the AECC: The Automotive Edge Computing Mobile World Live is the premier destination for news, insight Consortium (AECC) is an association of cross-industry, global and intelligence for the global mobile industry. Armed with a leaders working to explore the rapidly evolving and important dedicated team of experienced reporters from around the data and communications needs involved in instrumenting world, we are the industry’s most trusted media outlet for billions of vehicles worldwide. The AECC’s goal is to find more breaking news, special features, investigative reporting, and efficient ways to support the high-volume data and intelligent expert analysis of today’s biggest stories. services needed for distributed computing and network architecture and infrastructure. The AECC’s members are key We are firmly committed to delivering accurate, quality players in the automotive, high-speed mobile network, edge journalism to our readers through news articles, video computing, wireless technology, distributed computing, and broadcasts, live and digital events, and more. Our engaged artificial intelligence markets. audience of mobile, tech and telecom professionals, including C-suite executives, business decision makers and influencers For more information about the AECC and its depend on the unrivalled content and analysis Mobile World membership benefits, please visit aecc.org. Live provides to make informed business decisions every day.

Since 2016, Mobile World Live has also had a team of in-house media and marketing experts who work directly with our brand partners to produce bespoke content and deliver it to our audience in strategic yet innovative ways. Our portfolio of custom work - including whitepapers, webinars, live studio interviews, case studies, industry surveys and more – leverage the same level of industry knowledge and perspective that propels our newsroom.

Mobile World Live is published by, but editorially independent from, the GSMA, producing Show Daily publications for all GSMA events and Mobile World Live TV – the award-winning broadcast service of Mobile World Congress and home to GSMA event keynote presentations.

Find out more at www.mobileworldlive.com

1. GSMA: "The Connected Vehicle Opportunity," January 2021 (https://www.gsma.com/iot/wp-content/uploads/2021/01/Infographic-The-Connected-Vehicle-Opportunity.pdf)

2. McKinsey: "Unlocking the full life-cycle value from connected car data," February 2021 (https://www.mckinsey.com/industries/automotive-and-assembly/our-insights/unlocking- the-full-life-cycle-value-from-connected-car-data)

3. GSMA Intelligence, February 2021 (https://www.gsmaintelligence.com/data/)

4. SBD Automotive: "Connected Services Forecast," 2020 (https://www.sbdautomotive.com/en/connected-services-forecast-report)

5. M. Muehleisen, “Automotive Trials for Make-Before-Break 5G Edge Cloud Handover – Results and Takeaways from Joint Trials of Toyota, Vodafone and Ericsson at the Aldenhoven Testing Center,” 2020 (https://5g-ppp.eu/wp-content/uploads/2020/07/Maciej-Keynote-CLEEN-Workshop-Published.pdf)

10 WHITEPAPER