This post was originally written by Micheal Zimmerman, Bitfusion CEO on May 21, 2018
Edge computing is a well understood industry migration path. The need to push compute, storage and (obviously) networking to the edge is driven by IoT devices requiring better response time and locality of data. Many industry players are already innovating and designing software and hardware platforms tailored to fit the edge.
At Bitfusion we are exploring with our customers how they can benefit from ML/AI processing at the edge. In some use cases, there is a compelling value to perform inferences and some limited re-training at the edge.
This means that GPUs and FPGAs are needed, but the scale and dimensionality of the edge dictate the need for ML platforms at a non-data-center scale. Your typical data-center scale GPU will not fit. With Bitfusion’s partial GPUs and Network Attached GPUs, we provide optimized and scalable Elastic AI at the edge.
Please go to https://www.vmware.com/solutions/business-critical-apps/hardwareaccelerators-virtualization.html for more detailed information and collateral.