A typical Edge device node

So far, we have been using an Edge device node with Linux running on an amd64 CPU architecture. Using arm- or arm64-based CPU architecture Edge devices are equally good options. To keep the example simple, we will use CPU-based machine inferencing. However, in actual deployments, your Edge device nodes may be GPU-enabled. The process of using Open Horizon-based application workload deployment and ML model management is the same for all architectures. As noted in earlier chapters, whichever CPU or GPU architecture you choose, the host operating system should have a Docker runtime available and should support systemd.