Where will connected vehicle data flow and how will its lifecycle be managed? Freddie Holmes spoke with Dell to find out.
When a single connected vehicle roams in a smart city, it will generate and share a considerable amount of data. As vehicles add more features and become more autonomous, the amount of data generated per vehicle will increase significantly. Where this data goes, how it's used, and how long it's stored varies from application to application, but one thing is clear: data lifecycle management is becoming increasingly important to the automotive industry, and how to use these services will be critical.
Today, new cars are already highly connected, but will become smarter in the coming years, increasing the amount of data they generate and use. At the same time, cities are being equipped with 5G base stations, sensors inside multi-access edge computing (MEC) units, roadside units (RCUs), and compute and storage servers to help enable communication between a large number of different devices and workloads. In order to achieve a mutual dialogue between the car and the city, data must be shared quickly, reliably, and often in large quantities. As a result, edge computing has become a hot topic for automotive and smart city workers.
Scale your infrastructure
For example, edge servers process data collected from different connected devices around cities more tightly than cloud servers. By having a local edge server, upload and download speeds as well as latency have all been greatly improved, which will become invaluable in the coming years as more and more smart vehicles hit the road. Autonomous vehicles (AV) will generate higher data demands.
Connected, self-driving cars collect, process, and share data as they move through smart cities
In today's automotive industry, edge computing is primarily used in controlled test and development environments outside of the primary data center. As more and more connected vehicles enter public roads, the role of edge computing will rise.
"When we start talking about these vehicles going into production, the edge becomes very big and very fast," explains James Singer, a technologist in Dell Technologies' Infrastructure Solutions Group. "Edge infrastructure needs to send data where it needs to go; today, much of it goes to the cloud or on-premises data centers. Due to the expected amount of data, there will be many use cases that need to be processed at the edge. ”
For example, there may not be enough time or bandwidth to send blocks of data to a data center or cloud to clean, label, train, and send back to the car as an over-the-air (OTA) update. "Some kind of edge computing is needed to send data to a server close to where the vehicle is," Singer explains. "This infrastructure doesn't exist on a large scale yet, but it will exist as the industry continues to improve vehicle capabilities."
Singer said an important question the auto industry might ask or already be asking is how data is transferred from the vehicle to the cloud or colocation data centers. The industry is also working to determine who will guarantee that the data will arrive safely at its final destination, and who can see the data after it leaves the car. Then there's the question of whether the data remains on the Edge Server or whether all the data must be sent to one location. "Since the data is movable and decomposed, how will the data be collected and find the path back?" Singer asked. "Do automakers or Tier 1/Tier 2 suppliers need to build their own edge computing, networking, and storage? These are all valid questions. ”
To be sure, the automotive industry will consume compute and storage capabilities as services.
Edge as a service
Questions like these revolve around a key consideration: How will the automotive industry consume edge computing? Dell believes that Edge as a Service (EaaS) is the answer and is currently working on what is needed to realize this vision.
The idea is that EaaS will allow automakers to leverage the skills and resources of end-to-end edge infrastructure providers. In fact, it will provide turnkey edge solutions for automakers as they look to bring new connected and autonomous driving capabilities to the mass market. It follows similar trends around software-as-a-service (SaaS) and platform-as-a-service (PaaS), both of which have accelerated the industry's shift to digitalization.
Dell's technicians and engineers are working on a multitude of variables that affect how EaaS is delivered, from environmental factors such as power and cooling requirements to physical and data security, how compute and storage will be delivered, and whether some data remains in a particular location or eventually returns to the cloud. "There are still a lot of unknowns about how data will flow in the car assembly line," Singer emphasizes. "But one thing is for sure, the automotive industry will consume compute and storage capabilities as a service."
Automotive, edge, cloud
In theory, this kind of mobile sensor data would be put to good use, but as Singer explains, it's a priority question about which data leaves the car, which data stores, and even which data may no longer be needed. "In a mature production environment, we need to be more informed about which data leaves the car, which ones are deleted, which are stored, and which can be processed locally in the car," he says.
Recent advances around cellular connected vehicles (C-V2X) will further complicate work on data management. "There will be a lot of communication between C-V2X, RSUs, base stations, on-premises and cloud infrastructure, and other vehicles, and that's where data volumes start to become real challenges, and the 'conscious' surrounding resources are not only creators of data, but also consumers of data," he adds.
Managing high-resolution camera data will be a challenge, but Edge will prove to be a useful outlet
A vehicle's on-board sensors are constantly collecting information, for example, a vehicle may find that a collision has caused a roadblock, or a group of nearby school-age children may be in danger when approaching a crosswalk. A section of the road may be icy or damaged, and a safety driver or passenger (if fully autonomous) can send a distress signal to the emergency services. When video data starts to be shared in large volumes and with high quality, the edge becomes invaluable.
"Basic data transfer from vehicle to vehicle and from vehicle to infrastructure may be in the range of kilobytes," Singer explains. "But when we talk about the stream of data from cameras on these self-driving cars, that's going to accumulate a lot of data." Even if it's just 10 percent of the data generated by various sensors in autonomous vehicles, and then multiplied by thousands of cars, the scale will become enormous. "While lidar produces more data than radar, GPS, ultrasound, and IMUs, it doesn't come close to the amount of data coming from cameras. Singer observed that many vehicles already use 1K cameras, but this will transition to 4K and possibly even 8K in the future.
Singer stressed that pushing data back into the core data center will be "a long journey," so the overall trend in the coming years will be to position computing power and storage closer to where the data was generated. Since edge servers will be in a different environment than a secure core data center, the next challenge will be to ensure that this infrastructure does not fall prey to cyberattacks. "There are many potential attack vectors, and in this environment, the idea of a firewall will no longer be good enough," Singh warned.
This will require strong partnerships between many different companies to achieve this.
EaaS will handle data avalanches
As megatrends converge on 5G, autonomous driving, and smart cities, the end-to-end edge infrastructure available to automakers through EaaS will prove invaluable. This ecosystem will scale from the edge to core data centers and the cloud, helping to manage the increasingly complex data lifecycle of next-generation mobility.
The benefits of edge computing for the automotive industry and smart city developers are obvious, but a single player can't solve all of these problems on their own. Singer urges stakeholders across the ecosystem to work together to make this edge application a reality. "Dell won't be able to build all of this edge infrastructure on its own," he concludes, "and it will require strong partnerships between many different companies to achieve this." ”