Edge Computing Pushes Forward with Help From FPGAs to Testbed Environments
As more and more IoT devices become commonplace, companies like Microchip, IBM, Verizon, Stratus, and Schneider Electric are hoping to propel edge computing into the future.
Edge computing is amongst the hottest fields in the world of electrical engineering right now. With the advent of the Internet of Things (IoT) requiring low power and low latency performance, researchers and industry engineers alike have been working to enable AI computing on small, battery-powered edge devices.
A high-level overview of edge computing. Image used courtesy of Sridhar Sunkara and eBiz Solutions
Thanks to the momentum from the IoT boom, this field has been growing at an unprecedented pace, and a quick look at recent headlines signals that this pace is far from slowing down. This article will take a look at three recent industry developments in edge computing to see where the technology is and where it hopes to go.
Bring Computing Closer With Micro Data Centers
The first development comes in the form of data centers. Traditional, decentralized data centers are often less ideal than edge computing because of their remote locations, making it harder to support low latency and its public nature, utilized by thousands of customers simultaneously, inherently posing security concerns.
One solution, which admittedly is not “edge computing” but is undoubtedly tangential, is a micro data center, a small-scale computing infrastructure deployed on sight.
A Micro Data Center from Schneider and Stratus. Image used courtesy of Stratus
Coming in 6U, 12U, or 42U rack sizes, the micro data centers can best support 100,000 I/O’s and twenty remote clients. The company hopes that bringing the data center operations closer to the application will help solve data latency, bandwidth, and security issues related to traditional data centers.
Though this is just one piece of a solution to push computing to the edge, it is hoping to combat the growing issue of having too much data to compute. Another company that is also hoping to keep edge computing on a forward path is Microchip.
Low Power FPGAs from Microchip
A more traditional approach to edge computing is to cut down the power consumption and thermal profile of computing hardware to support the requirements of edge computing.
The company’s new offerings, the PolarFire FPGA and PolarFire SoC FPGA claim to have cut power consumption by as much as 50% compared to the previous generation, a large portion of which is attributed to lower static power consumption.
While maintaining high performance, the new FPGA can leverage this decreased power consumption to cut its footprint in half, reaching as small as 11mm x 11mm.
The PolarFire SoC FPGA can reach a footprint as small as 11mm x 11mm. Image from Microchip
The FPGA is built on a RISC-V architecture-based processor, including two megabytes of L2 cache, LPDDR4 support, and up to 50K logic elements. Overall, Microchip claims they are intended for applications including embedded computer vision, security, and general machine learning.
Microchips solution for edge computing was a more direct solution than micro data centers; however, they both have their specific roles to play in this field. One final advancement for edge computing comes from IBM and Verizon.
IBM and Verizon Team Up to Create a Testbed
Finally, the last piece of edge computing news centers around a collaboration between IBM and Verizon.
Last week, the companies announced that they are creating a testbed environment at IBM’s Industry Solution Lab in Texas, which designers will use explicitly to develop 5G-enabled edge applications.
The idea is to allow enterprise customers at the lab to develop and test edge devices leveraging Verizon’s 5G network, which is still in the process of being rolled out.
From IBM’s perspective, it believes that the testbed could provide a pre-commercial, standalone 5G and multi-access edge computing (MEC) environment. Also, IBM hopes that it could also help provide the technical resources and expertise needed to test and optimize products.
The company says that the development efforts will focus initially on asset monitoring and optimization, worker productivity and safety, and visual inspection, all of which are relevant industry 4.0 applications.
Although this testbed is a longer-term goal, hopefully, the benefits of testing and developing more edge computing devices can help further edge computing to the next level in the near future.
A Growing Field
Bringing computing closer to the physical application is a must in today’s world, and many people are working very hard to make this a reality.
Based on the variety of our headlines, it’s clear that companies are approaching this challenge in different ways, but what’s important is that progress is being made in whichever direction they choose.
Between IBM, Verizon, Microchip, Stratus, and Schneider Electric, there’s clearly no lack of innovation in the field. It will be interesting to see how these companies, and more, aim to solve the challenges of pushing computing to the edge.