By KV Thanjavur Bhaaskar, IIoT & AI Solutions Architect at Xilinx
Let’s talk about data. The amount of data produced is exploding and our existing infrastructure is struggling to keep up. We are seeing industrial, vision, and healthcare equipment manufacturers getting out in front of this trend by working more closely with Xilinx. Why, you ask? Smart factory, surveillance, and hospital assets contribute heavily to the data explosion, and these companies and their customers want to figure out a way to manage data economically and securely. Xilinx’s heterogeneous and scalable SoCs address today’s and tomorrow’s needs!
Economically, there is the cost of slow response and the dollar cost. If we use the example of an industrial control system on the east coast of the United States that has a (fairly typical number) 10ms control cycle that pings a data center on the west coast of the US, the response will come back in ~30ms (Laws of Physics). This won’t meet the needs of that system, so we must do it locally (Edge Intelligence). Next is the dollar cost challenge. If we send large amounts of data to the cloud, we end up paying for transmission, computation, and storage of the data. Alternatively, we can preprocess the data locally and send only aggregated, filtered, or meta data to the cloud (Edge-Cloud Cooperation), saving substantial amounts of money! Furthermore, pre-processing data obfuscates it and enhances privacy. Xilinx supports best-in-class cybersecurity standards such as IEC 62443. You can read about it in this recently released Xilinx white paper.
The Industrial and Healthcare IoT team has created multiple solutions to address Edge Intelligence and Edge-Cloud Cooperation, let’s highlight two of them here.
Edge Intelligence Solution
With our “SPYN” edge intelligence kit, we demonstrate industrial control, software upgradability, python programmability, visualization, data manipulation, and finally predictive maintenance—all done at the edge (near the analog-digital boundary). Therefore, the Electric Drive Demonstration Platform (EDDP) was born, which shows developers how to implement their motor control algorithm on Xilinx adaptable and scalable SoCs using our tools. The kit can be upgraded to SPYN for free with a simple Ethernet cable and an Internet connection. The upgrade brings you into the world of python, where all open source python libraries are fair game. The final free upgrade of the kit to SPYN AI (coming soon!) enables machine learning inference at the edge, which performs predictive maintenance and alerts the user on the health of the motor.
Edge-Cloud Cooperation Solution
Edge-cloud cooperation was covered in detail by Wes Skeffington (Principle Architect at Xilinx) in his previous blog post and solution GitHub.
Now let us show you!
We have been conducting a series of free workshops to provide developers first-hand experience of the solutions.