Red Hat Joins Facebook’s Open Compute Project

Red Hat has joined Facebook’s project to make a greater use of open source hardware within new data centres

Open-source giant Red Hat has added its backing of open source in data centers, after it joined Facebook’s Open Compute Project.

Red Hat’s first step as a member of the Open Compute Project was to certify its Red Hat Enterprise Linux operating system to run on two servers based on the project’s specifications, the open-source software giant said 27 October.

After certification is complete, Red Hat is expected to test its Red Hat Enterprise Virtualisation platform and storage technologies it obtained as part of the Gluster acquisition on those servers.

Open Compute Project

“With the Open Compute Project, Facebook is using the power of the open-source model to redefine how data centres are designed and built. With our consistent focus on innovation and end-user value, it was natural to take part in this project to not only extend the benefits of Red Hat products to these systems, but to also help expand the reach of open-source technologies as a whole,” said Brian Stevens, CTO and vice president of engineering at Red Hat.

Facebook launched the Open Compute Project in April after building a highly efficient data centre in Prineville, Oregon. The data centre is considered the most efficient in the world in terms of power consumption, using 38 percent less energy than the company’s other data centres while costing 24 percent less. The power usage effectiveness rating is 1.07, and only 7 percent of the power brought into the facility is used to cool down the facility.

Building the facility required Facebook engineers to custom-design servers and all the server-room hardware, including power supplies, battery backup systems and racks, to meet their requirements. The company launched Open Compute Project and published some of specifications and designs of the hardware developed for Prineville. The aim was to encourage collaboration in designing hardware and systems efficient enough for large and powerful web data centres by adopting the model used in open-source software development.

By sharing its hardware specifications with other companies, Facebook can get suggestions on how to improve the servers faster than if the company’s engineers studied the systems on their own. Outsiders may spot inefficiencies or add specific features that may help certain industries and improve the hardware before internal engineers would have gotten around to implementing them.

Audacious Idea

Dell, Intel, Advanced Micro Devices and Asus have already contributed intellectual property to the project, including motherboard and blade specifications. For example, Intel and Facebook collaborated on the designs of two Intel motherboards that met Open Compute Project specifications.

“What began a few short months ago as an audacious idea – what if hardware were open? – is now a fully formed industry initiative, with a clear vision, a strong base to build from and significant momentum,” Frank Frankovsky, Facebook’s director of hardware design and supply chain, wrote on the Open Compute Project’s blog.

Frankovsky also announced a new foundation that will guide the Open Compute Project at the company’s Open Compute Summit in New York City on 27 October. The Open Compute Foundation will drive collaboration efforts, Frankovsky said.

“We need to continue to grow the community and enable it to take on new challenges. We need to ensure that, as the community evolves, it retains its flat structure and its merit-based approach to evaluating potential projects. And we need to keep the community focused on delivering tangible results,” Frankovsky wrote.

Facebook has another data centre in North Carolina and at the Open Compute Summit announced it will build its first European data centre in Lulea, Sweden. The facility is expected to go live in 2014 and will be three times the size of the Prineville facility.