Verge.io, formerly known as Yottabyte, has added GPU virtualization to its virtual data-center software, making it possible to partition out GPU-based processing tasks the same way a hypervisor divides up a CPU.
Hyperconverged infrastructure (HCI) builds a tightly integrated and configured IT data center setup from standard server, storage, and networking servers with everything virtualized. Verge.io’s software, called Verge-OS, goes on step further by partitioning the pooled resources into what it calls virtual data centers (VDCs). From one big pool can come many VDCs made up of different hardware and software configurations.
Once Verge-OS is installed it pools hardware assets—“liquefies” them as CEO Yan Ness puts it. That pool of CPUs, memory, storage, and now GPUs can be apportioned into individual virtual data centers. These VDCs are independent of each other and can be moved around the physical hardware or physical data center locations.
The VDCs can run a single app or multiple apps and can be saved as blueprints for immediate loading of configurations, with the VDCs switchable on the fly. So for example an enterprise data center might be configured to run line-of-business apps during the work day, then at night when no one is using the network, a VDC for high performance computing tasks could take over.
With the GPU support in Verge-OS, customers are already using GPUs for remote desktops during the day and at night are running ML and AI workloads on it. “[GPUs] are just another liquified resources that you can assign to a VM or a virtual data center anywhere in the stack,” said Ness.
Back in its Yottabyte days, the company was focused on software defined storage and software defined networking. It expanded its offerings to include a hypervisor, backup and recovery, system management and integration tools. The main selling point of Verge-OS was a customer could get all that software from one source rather than many. The vGPU is just the latest piece of hardware it is supporting in the VDC.
Verge.io does not demand the most bleeding edge of hardware to run its virtualization software. “Give us commodity hardware. We have demos that run on laptops,” said Ness.
To use the Nvidia GPUs with Verge-OS you need what is known as a vGPU license from Nvidia, which lets you use their GPUs in a cloud or virtual setting.
And Ness said the software is self-monitoring, self-healing, and self-optimizing technology. “So you don’t need all these storage engineers and whatever to make sure that your settings are appropriate for the workloads that you have,” he said.
Ness said Verge-OS will support AMD Instinct cards in the future, it’s just that Nvidia is so dominant Verge went with it first. But AMD will be supported in the VDC software partitioning, eventually. “We want to do bit more testing,” he said.
Verge-OS with vGPU support is available from Verge.io now.