At VirtualSC20, we visited with Jaan Mannik of One Stop Systems to talk about the company’s “AI on the Fly” strategy. Mannik, who is OSS director of western U.S. & international sales, says AI on the Fly takes on the technical challenge of bringing the power of data center to the edge for advanced compute- and data-intensive workloads.
We basically take bleeding edge GPU compute, and flash storage hardware, and we put it out of the traditional uses of a data center and closer to edge applications, whether it be airplane surveillance, autonomous driving, media and entertainment. A lot of these applications struggle with getting the performance that they require at the edge.
Mannik elaborates on the requirements needed to support its high performance edge mission, a range of technologies encompassing more than GPUs and flash arrays.
As these edge applications become more complex – and it’s not just with the resources that we’re using from companies like Nvidia – this is also taking advantage of the high-speed design core competency of One Stop Systems and wrapping it in a specialized type or form factor that can take advantage of the different power requirements, the different type of milspec (military specification) organizations, really about getting the power of the data center to the edge.
