Distributed computing in a local network.
Vehicles connected to the network will process and produce huge amounts of data, and whoever has access to this data will be able to benefit enormously from it.
If data is indeed the new oil, then those players who are trying to get exabytes of data from connected vehicles will speculate on it. Toyota wants to be one of those players.
Of course, OEMs should be the main beneficiaries of vehicle-generated data. Following Tesla's model, many automakers are looking to ditch one-off car sales and move to pegging buyers through the sale of over-the-air upgrade subscriptions. With the many applications and services available for connected vehicles, mobile operators are also capitalizing on the imminent and significant growth in traffic. Cloud service providers and tech companies are also interested in these processes as they strive to meet the growing demand for distributed computing.
At least that's how it all works in theory. This state of affairs motivates at least one OEM to be involved in the Advanced Automotive Technology Consortium (AECC): Toyota. AECC (Automotive Edge Computing Consortium) is a non-profit cross-industry group. The AECC was founded at the 2017 Mobile World Congress and is actively defining a set of network requirements for connected vehicles. The consortium includes Toyota, Ericsson, Intel, Denso and NT&T. AECC sponsors are Cisco, Dell, KDDI and Samsung.
The AECC does not focus on V2X (vehicle-to-vehicle and vehicle-to-infrastructure communication) or networking in infotainment systems. Instead, the group focused on developing specifications to maximize the collection and processing of big data generated by connected vehicles.
The AECC has published the second version of its technical report Driving Data to the Edge: The Challenge of Data Traffic Distribution . It can be downloaded for free from this link.
Big Data Challenges
In an interview with the EE Times, Kenichi Murata, President and Chairman of AECC and General Manager of Projects at Toyota Motor, said: "The industry initially assumed that in 5 years, each vehicle would generate about 1 gigabyte of data per month." By now, the AECC has realized its underestimation. With a wave of new services and applications emerging, the AECC is forecasting data generated to grow to 1-10 exabytes of data per month in 2025.
This exponential growth can be tackled by creating a localized / metropolitan network and defining where data processing and computation should occur on the “edge” of the network.
The key to this solution is the localized network. It's better to analyze the data locally because it doesn't waste network bandwidth and doesn't have to be prohibitively expensive to transfer, explained Leifeng Ruan, chairman of AECC WG2 and chief engineer at Intel.
The AECC White Paper 2.0 provides specific solutions and recommendations for offloading data at the edge of the network, choosing servers and mobility providers, and making automotive systems available.
Strategy for cellular operators
The networks referred to by the AECC are cellular networks. The latest technical article published by the AECC provides detailed guidance for mobile operators interested in dealing with the growing demand for data from connected vehicles. The group noted, "For carriers, the report provides information on the functionality required by next-generation networked vehicles, as well as recommendations for network configurations to support the emerging transport network ecosystem."
Murata noted that the AECC has discussed its progress with 3GPP (a consortium that develops specifications for mobile telephony), although the list of cellular operators included in the working group includes only Japanese NTT, KDII and AT&T.
In its technical paper, the AECC also highlights the key roles that are required of "distributed data management" in a "topology-oriented distributed cloud architecture". This technology is also known as "edge computing".
The list of companies providing cloud services at the AECC includes Oracle, Dell, Google Cloud and Microsoft Azure.
The basic need for networked vehicles for distributed networks and local data integration resembles the challenge that Netflix and YouTube were solving to deliver content to client devices. For boot services to work efficiently, they needed an edge network with a caching mechanism. The specificity of working with connected vehicles is that vehicles not only upload large amounts of data from the cloud, but also upload them there, Murata explained.
What about OEM car manufacturers?
While the AECC has worked hard on its technical recommendations for a network architecture capable of handling large amounts of data from connected cars, the industry consortium desperately needs the involvement of other automakers besides Toyota.
Murata acknowledged the problem. But Toyota is confident that other players in the market (especially smaller OEMs) will follow suit.
What about other major automakers in Europe, for example? BMW, VW and Daimler have said they are tackling big data challenges with cellular operators through the 5G Automotive Association (5GAA). Murata comments on this situation as follows: "This is not true, I know for sure."
As it turns out, 5GAA is paying more attention to other issues, in particular the cellular-based V2X.
Sooner or later, the problem of working with big data must be solved at the intersectoral level, Murata explained. With a focus on the collaboration between AECC and 3GPP, Murata hopes that 3GPP can act as a bridge to bridge the gap with 5GAA.
The AECC chairman said the devil is in the technical details of how mobile operators, automakers and cloud service providers interact. The LAN must provide support for distributed computing, on-demand data tracking, and data storage and indexing, and the types of data sent by different vehicles are likely to be different. The chairman also added that "there are a number of technical elements that need to be standardized."
Toyota and connectivity
Be that as it may, the results of the work of the AECC over the past few years have not received wide publicity.
Egil Juliussen, an auto industry analyst, noted that automakers are inherently reluctant to share data among themselves. If big data is an oil well, why drill with anyone?
Juliussen noted that Toyota, which is a bit late to the connectivity market (compared to GM with their On Star or BMW with Connected Drive), may be looking to catch up in hopes of getting ahead of the rest.
Juliussen argues that as with any IoT project, the challenge of working with big data boils down to two questions. First, who is going to pay for networking? Second, what business model can truly leverage connectivity?
Speaking about the prospects for automotive architecture in the next 5-10 years, Juliussen believes that the edge computing technologies needed to process data will not be embedded in the network infrastructure, but in the cars themselves.
Automakers and tech companies are focused more on reducing the amount of data sent to the cloud. Murata objected: he believes that reducing the volume of downloaded data is a problem today. In turn, he noted that the AECC is working on "the challenges of the day of tomorrow", that is, a next generation network architecture capable of meeting the growing demand for applications and services for connected vehicles.
Subscribe to the channels:
@TeslaHackers - a community of Russian Tesla hackers, rental and drift training on Tesla
@AutomotiveRu - auto industry news, hardware and driving psychology
About ITELMA
Read more helpful articles: