Connected vehicles will consume and produce massive amounts of data. Having possession of that data could be incredibly lucrative.
If data is indeed the new oil, then the new wildcatters will be those trying to tap the exabytes of data from connected vehicles. Toyota wants to be one of them.
Naturally, car OEMs stand to be the primary beneficiaries of data generated by connected vehicles. Following the Tesla model, many automakers want to transform from one-off vehicle sales to tethering car buyers with subscriptions for regular over-the-air updates. With a plethora of connected vehicle applications and services on the horizon, mobile network operators are spoiling to also benefit from the imminent massive increase in data traffic. Cloud service providers and tech companies also want in on the action, so they can meet the growing demand for distributed computing.
At least that’s the theory. It’s motivating at least one big car OEM, Toyota, to put muscle behind the Automotive Edge Computing Consortium (AECC), a non-profit cross-industry group. Founded at the Mobile World Congress in 2017, the AECC has been actively charting the course of network requirements for connected vehicles. The consortium’s charter members include Toyota, Ericsson, Intel, Denso and AT&T. Its sponsored members are Cisco, Dell, KDDI and Samsung.
The focus of the AECC is neither V2X (vehicle to vehicle, or vehicle to infrastructure) communication nor connectivity required for vehicle infotainment. Instead, the group has zeroed in on the technical requirements to best collect and process the big data generated by connected vehicles themselves.
The AECC published Wednesday Version 2.0 of its technical report, Driving Data to the Edge: The Challenge of Data Traffic Distribution. It is available for download at no charge by visiting here.
The big data challenge
In an interview with EE Times, Kenichi Murata, president and chairman of the AECC, and Toyota Motor’s general project manager, said, “The industry originally assumed that each vehicle would generate something like 1 Gigabyte data per month — five years from now.” By now, however, the AECC has realized its underestimate. Given the wave of emerging services, the AECC projects data volume to shoot up to 1-10 exabytes of data per month in 2025.
The way to deal with such exponential growth is to create a localized/metro network and identify at which “edge” of the network data processing/computation should take place.
The key here is the localized network. You want to analyze data locally, because you shouldn’t waste network bandwidth and data transmission shouldn’t be cost prohibitive, explained Leifeng Ruan, AECC WG2 chair and Intel principal engineer.
The AECC’s Version 2.0 tech paper lays out specific solutions and recommendations for edge data offloading, mobility service provider server selection and vehicle system reachability.
Playbook for cellular operators
The communication network the AECC refers to here is cellular networks. With this newest tech paper, the AECC has effectively written a playbook for mobile network operators (MNOs) who choose to serve the growing data demand from connected vehicles. The group noted, “For MNOs, the report provides information on the functionality that is required by the next generation of connected cars and recommendations on how their networks must be configured to support the emerging connected car ecosystem.”
Murata noted that the AECC has been in touch with 3GPP (The 3rd Generation Partnership Projec) about AECC progress, although cellular network operators listed in the industry group today are limited to Japan’s NTT, KDDI, and AT&T.
The AECC also highlights in its tech paper the key roles that must be played by “data management in a distributed manner” in its “topology-aware distributed clouds architecture,” aka edge computing.
Cloud service companies in the AECC include Oracle, Dell, Google Cloud and Microsoft’s Azure.
Connected vehicles’ basic need for distributed network and local data integration is reminiscent of what Netflix or YouTube demanded for delivering content to client devices. For efficient downloading services, they needed an edge network with a caching mechanism. The difference with connected vehicles is that vehicles are both downloading data and uploading massive amounts of data to the cloud, explained Murata.
What about other car OEMs?
While the AECC has worked diligently on its technical recommendations for network architecture capable of handling connected vehicles’ big data, the industry consortium sorely needs participation from car OEMs other than Toyota.
Murata acknowledged the problem. But Toyota is confident that others — especially smaller OEMs — will follow suit.
What about other big carmakers, perhaps in Europe? BMW, VW and Daimler have said that they’re tackling the big data issues together with cellular operators through the 5G Automotive Association. But “that’s not true,” stressed Murata. “I know this for fact.”
As it turns out, 5GAA has been more focused on such issues as cellular-based V2X (vehicle to vehicle, vehicle to infrastructure).
Sooner or later, the big data challenge must be analyzed, on a cross-industry level, Murata explained. Stressing the AECC’s relationship with the 3GPP, Murata is hopeful that 3GPP can serve as a liaison to bridge the gap with 5GAA.
However, the devils are in the details of the technical issues among mobile network operators, car OEMs and cloud service providers, explained the AECC chairman. The local network must be able to support distributed computing, track data upon request, save data and index data at a time when not every vehicle is likely to send the same types of data. “There are technical elements that must be standardized,” he said.
Toyota and connectivity
What the AECC has been doing for the last few years, however, hasn’t been broadly publicized.
Egil Juliussen, an automotive industry analyst, pointed out that automakers by nature tend not to share data among themselves. If big data is the next bog gusher, why invite others to the drilling?
Juliussen observed that Toyota, somewhat late to the connectivity game compared to GM (with On Star) or BMW (Connected Drive), might be overcompensating this time around, in hopes of getting ahead of the crowd.
As with any IoT project, the big data issue comes down to two issues, said Juliussen. First, who is going to pay for connectivity? Second, what’s the business model that can truly leverage connectivity?
Thinking about automotive architecture for the next 5-10 years, Juliussen believes the edge needed for data processing won’t be necessarily at the edge of the network but in vehicles themselves.
Carmakers and tech companies are more intent on reducing the data that must be sent to the cloud. Murata, however, countered, “Reducing the data” that must be uploaded to the cloud is “today’s problem.” In contrast, he noted, the AECC is working on “tomorrow’s problem,” next-generation network architecture that can meet the growing demand for connected vehicle apps and services.