AMD’s SeaMicro Microservers Will Drive Verizon’s New Cloud
The SeaMicro SM15000 systems and Freedom fabric will be the drivers of these capabilities, Feldman said. When SeaMicro initially was developing its microservers, they were based on Intel’s low-power Atom platform. However, they now also use server chips from AMD (Opteron) and Intel (Xeon), and the SM15000s that Verizon will run will use all three chips. The microservers are designed to pack a lot of performance into small systems that use relatively small amounts of power and space, enabling businesses to put a lot of them into data centers. The SM15000s will run up to 512 cores in a 10U (17.5-inch) system, with up to 2,048 cores per rack and 4 terabytes of memory per system. Each CPU socket offers 10G-bps bandwidth, and the systems bring up to 5 petabytes of storage. A key is the Freedom fabric, which Feldman said brings a greater level of security and reliability to Verizon’s public cloud. Cloud providers traditionally leverage the multitenancy concept, where workloads run on the same servers and use the same networking connections. Performance can be hurt if demand is too high on the systems and networks, creating a bottleneck. Via the Freedom fabric, workloads are assigned dedicated pathways to avoid the bottleneck, ensuring that there is no commingling of the jobs, Feldman said.The Verizon deal is part of a larger effort to bring AMD back to profitability and reduce its reliance on the contracting global PC market. Low-power dense servers for cloud and other hyperscale environments are one of several growth areas that AMD is targeting, with the others including ultraportable devices, embedded systems and the company’s semi-custom business. The public cloud space promises to be a booming one for the next several years, according to IDC analysts. In a report last month, IDC said the public IT cloud server market will grow 23.5 percent between now and 2017, with spending growing from $47.4 billion in 2013 to more than $107 billion in four years.
Greater user control also is factor, he said. By pooling and aggregating resources, businesses can more easily determine the size of capabilities of the virtual machine. If a customer wants a VM with a lot of memory or more performance or less storage, that can be easily done.