LinkedIn donates its Open19 specs to the Open Compute Project

Open19 creates an open standard for server, storage and networking devices so that they fit into a 19-inch rack. (Getty Images)(data center)

LinkedIn has contributed its Open19 specifications to the Open Compute Project, making the specs available to the OCP community.

Open19 is set of data center specs aimed at enabling hyperscale growth of data centers, according to a LinkedIn post by Zaid Ali Kahn, LinkedIn's head of infrastructure engineering. The initial mandate was to address data centers of all sizes and create an open standard for server, storage and networking devices so that they fit into a 19-inch rack environment. The standard was to support high compute node density and disaggregation.

The OCP organization, Open19 and the data center industry more generally are dealing with massive growth in demand. At the same time, applications are moving to the cloud and the edge. In short, legacy data centers are being disaggregated. Data center technology now must fit into tighter, and in almost all cases far more challenging, spaces. Doing so in an open and repeatable manner is vital.

Open19 supports all 19-inch environments and full rack and software/hardware disaggregated technology. It is fully redundant across all element standards, offers a universal power solution, increases rack integration, simplifies server integration, provides high power conversion efficiency and improved power usage effectiveness (PUE) and is cost effective.

LinkedIn is benefiting by using Open19. "To further put this in context, since the Open19 tech has been deployed at LinkedIn, we have experienced up to 6X improvement in rack integration and now have the ability to install 4X the number of servers per rack," wrote Kahn.

RELATED: Microsoft splashes compression algorithm and hardware specs for Azure into open source

Open19 supports Project Olympus, which was developed by Microsoft and contributed to OCP in 2017. This, Kahn says, provides Open19 with a cloud-based, high-end Olympus-based server that support high memory and high core count CPUs.

LinkedIn joined OCP last May.