Microsoft has updated the cloud server specifications that it contributes to the Open Compute Project (OCP) with a new release incorporating the latest Intel chips, advanced networking and greater flexibility to boost performance relative to the cost of ownership.
The firm announced its second-generation Open CloudServer (OCS v2) specifications at the OCP European Summit in Paris, claiming that these deliver a number of improvements, particularly in terms of performance and flexibility.
The OCP was launched by Facebook to drive the development of more efficient servers and other hardware better suited to the scale and dynamic nature of modern large-scale data centre operations.
Microsoft announced in February that it had joined the OCP and contributed its original cloud server specifications.
As well as hardware specifications, this included design collateral such as CAD and Gerber files, plus system management source code used to deploy cloud services such as Windows Azure and Office 365.
Microsoft is now updating the specifications to support a diverse range of cloud services, as well as varying requirements around the globe that affect power and voltage to the rack, as well as safety and regulatory certifications affecting weight, rack size, and mechanical and cooling parameters.
The new blueprints specify a dual-processor design based on Intel's recently launched Xeon E5-2600 v3 processor family with support for 128GB, 192GB and 256GB memory capacity configurations, plus provision for M.2 flash storage, which enables servers to be fitted with a higher density of high performance solid state drives using the PCI Express bus.
Also included in the specifications is 40Gbps Ethernet connectivity with support for routable remote direct memory access (RDMA) over Converged Ethernet. This latter extends the RDMA capability, wherein data can be moved directly from the memory of one server node to another, so that such transfers can be routed over an IP connection.
Microsoft said it has also provided for the integration of a variety of components and add-on cards, including custom logic FPGA accelerators, enabling customers to tune their server performance for unique workloads.
Writing on Microsoft's Server & Cloud Blog, Kushagra Vaid, general manager for server engineering in Microsoft's Cloud & Enterprise division, said that the OCS v2 specification is the convergence of a number of design points.
"Meeting these design points in one server is critical for Microsoft to meet its service delivery and data centre operations goals, and also the needs of the community who'll use OCS v2 for their own data centre operations," he explained.
In particular, OCS v2 delivers the high performance required for Microsoft's own services, while minimising the overall total cost of ownership, according to Vaid.
"Reaching the sweet spot for a number of server performance metrics was critical for OCS v2 to be deployed widely across Microsoft's own global data centre footprint," he said.
Moon's dark side is mountainous, rugged and never visible from the Earth
The groundwater basins in some areas of Tehran have been damaged irreversibly
This is the first time that any spacecraft on Mars has recorded air vibrations on the planet
Arctic sea ice is thickening at a faster rate during winter, thus slowing down long-term decline: NASA
But, the seasonal ice growth could only delay the demise of the Arctic ice cap for a few more decades