Edge Computing is Here. What's Next?
December 09, 2020
Edge computing has not fully evolved from theoretical and experimental to being deployed at scale. In this paper, we discuss how leveraging separation kernel hypervisors can enable the secure and...
Consolidating multiple systems onto single multicore processors can reduce cost, power, and footprint, but it can also create real-time performance challenges due to sharing of resources. There is also a desire to lower costs and improve flexibility by transferring traditional industrial network functionality from hardware to software implementations. There is, however, an opportunity to address each of these issues by leveraging separation kernel hypervisors to enable the secure and safe sharing of multicore processors, and by redefining the second layer of the automation pyramid—the supervisory layer—to be software driven. Referencing the real-world example of a car manufacturer dramatically improving their quality-control processes, we will examine the results from a testbed created in association with a number of pioneering European companies that can deploy data analytics workloads, communication protocol translation, and control workloads on the same platform without compromising performance. We will also discuss specific performance measurements for a Virtual PLC architecture which validates that this is a feasible path for the industry. Finally, we will demonstrate that the combination of these technologies can help shift Industry 4.0 and edge computing from the experimental to the everyday