 | |
| Abbreviation | OCP |
|---|---|
| Formation | 2011; 14 years ago (2011) |
| Type | Industry trade group |
| Purpose | Sharing designs ofdata center products |
Region | Worldwide |
| Membership | 50+ corporations |
| Website | opencompute |
TheOpen Compute Project (OCP) is an organization that facilitates the sharing ofdata center product designs and industry best practices among companies.[1][2] Founded in 2011, OCP has significantly influenced the design and operation of large-scale computing facilities worldwide.[1]
As of February 2025, over 400 companies across the world are members of OCP, includingArm,Meta,IBM,Wiwynn,Intel,Nokia,Google,Microsoft,Seagate Technology,Dell,Rackspace,Hewlett Packard Enterprise,NVIDIA,Cisco,Goldman Sachs,Fidelity,Lenovo,Accton Technology Corporation andAlibaba Group.[1][3][2]
The Open Compute Project Foundation is a501(c)(6) non-profit incorporated in the state of Delaware, United States. OCP has multiple committees, including the board of directors, advisory board and steering committee to govern its operations.
As of July 2020, there are seven members who serve on the board of directors which is made up of one individual member and six organizational members. Mark Roenigk (Facebook) is the Foundation's president and chairman.Andy Bechtolsheim is the individual member. In addition to Mark Roenigk who representsFacebook, other organizations on the Open Compute board of directors includeIntel (Rebecca Weekly),Microsoft (Kushagra Vaid),Google (Partha Ranganathan), andRackspace (Jim Hawkins).[4]
A current list of members can be found on theopencompute.org website.
The Open Compute Project began in Facebook as an internal project in 2009 called "Project Freedom". The hardware designs and engineering team were led by Amir Michael (Manager, Hardware Design)[5][6][7] and sponsored by Jonathan Heiliger (VP, Technical Operations) and Frank Frankovsky (Director, Hardware Design and Infrastructure). The three would later open source the designs of Project Freedom and co-found the Open Compute Project.[8][9] The project was announced at a press event at Facebook's headquarters in Palo Alto on April 7, 2011.[10]
The Open Compute Project Foundation maintains a number of OCP projects, such as:
Two years after the Open Compute Project had started, with regards to a more modular server design, it was admitted that "the new design is still a long way from live data centers".[11] However, some aspects published were used in Facebook's Prineville data center to improve energy efficiency, as measured by thepower usage effectiveness index defined byThe Green Grid.[12]
Efforts to advance server compute node designs included one forIntel processors and one forAMD processors. In 2013,Calxeda contributed a design withARM architecture processors.[13] Since then, several generations of OCP server designs have been deployed: Wildcat (Intel), Spitfire (AMD), Windmill (Intel E5-2600), Watermark (AMD), Winterfell (Intel E5-2600 v2) and Leopard (Intel E5-2600 v3).[14][15]
OCP Accelerator Module (OAM) is a design specification for hardware architectures that implement artificial intelligence systems that require high module-to-module bandwidth.[16]
OAM is used in some of AMD'sInstinct accelerator modules.
The designs for a mechanical mounting system have been published, so that open racks have the same outside width (600 mm) and depth as standard19-inch racks, but are designed to mount wider chassis with a 537 mm width (21 inches). This allows more equipment to fit in the same volume and improves air flow. Compute chassis sizes are defined in multiples of anOpenU or OU, which is 48 mm, slightly taller than the typical 44mmrack unit. The most current base mechanical specifications were defined and published by Meta as the Open Rack V3 Base Specification in 2022, with significant contributions fromGoogle andRittal.[17]
At the time the base specification was released, Meta also defined in greater depth the specifications for therectifiers and power shelf.[18][19] Specifications for the power monitoring interface (PMI), a communications interface enabling upstream communications between the rectifiers andbattery backup unit(BBU) were published by Meta that same year, withDelta Electronics as the main technical contributor to the BBU spec.[20]
Since 2022 however, the power demands ofAI in the data center has necessitated higher power requirements in order to fulfill the heavy power demands of newerdata center processors that have since been released. Meta is currently in the process of updating its Open Rack v3 rectifier, power shelf, battery backup and power management interface specifications to account for these new more powerful AI architectures being used.
In May 2024, at an Open Compute regional summit, Meta and Rittal outlined their plans for development of their High Power Rack (HPR) ecosystem in conjunction with rack, power and cable partners, increasing the power capacity in the rack to 92 kilowatts or more of power, enabling the higherpower needs of the latest generation of processors.[21] At the same meeting, Delta Electronics andAdvanced Energy introduced their progress in developing new Open Compute standards specifying power shelf and rectifier designs for these HPR applications.[22] Rittal also outlined their collaboration with Meta in designing airflow containment,busbar designs andgrounding schemes to the new HPR requirements.[23]
Open Vault storage building blocks offer high disk densities, with 30 drives in a 2UOpen Rack chassis designed for easydisk drive replacement. The 3.5 inch disks are stored in two drawers, five across and three deep in each drawer, with connections viaserial attached SCSI.[24] This storage is also called Knox, and there is also a cold storage variant where idle disks power down to reduce energy consumption.[25] Another design concept was contributed by Hyve Solutions, a division ofSynnex in 2012.[26][27] At the OCP Summit 2016 Facebook together with Taiwanese ODM Wistron's spin-off Wiwynn introduced Lightning, a flexible NVMe JBOF (just a bunch of flash), based on the existing Open Vault (Knox) design.[28][29]
The OCP has published data center designs for energy efficiency. These include power distribution at 277 VAC, which eliminates one transformer stage in typical data centers, a single voltage (12.5 VDC) power supply designed to work with 277 VAC input, and 48 VDC battery backup.[12]
On May 8, 2013, an effort to define an opennetwork switch was announced.[30] The plan was to allow Facebook to load its ownoperating system software onto the switch. Press reports predicted that more expensive and higher-performance switches would continue to be popular, while less expensive products treated more like acommodity (using thebuzzword "top-of-rack") might adopt the proposal.[31]
The first attempt at an open networking switch by Facebook was designed together with Taiwanese ODMAccton usingBroadcom Trident II chip and is called Wedge, the Linux OS that it runs is called FBOSS.[32][33][34] Later switch contributions include "6-pack" and Wedge-100, based on Broadcom Tomahawk chips.[35] Similar switch hardware designs have been contributed by:Accton Technology Corporation (and its Edgecore Networks subsidiary), Mellanox Technologies, Interface Masters Technologies, Agema Systems.[36] Capable of runningOpen Network Install Environment (ONIE)-compatiblenetwork operating systems such asCumulus Linux, Switch Light OS by Big Switch Networks, or PICOS byPica8.[37] A similar project for a custom switch for theGoogle platform had been rumored, and evolved to use theOpenFlow protocol.[38][39]
Sub-project forMezzanine (NIC) OCP NIC 3.0 specification 1v00 was released in late 2019 establishing 3 form factors: SFF, TSFF, and LFF .[40][41]
In March, 2015,[42] BladeRoom Group Limited and Bripco (UK) Limited sued Facebook, Emerson Electric Co. and others alleging that Facebook has disclosed BladeRoom and Bripco's trade secrets for prefabricated data centers in the Open Compute Project.[43] Facebook petitioned for the lawsuit to be dismissed,[44] but this was rejected in 2017.[45] A confidential mid-trial settlement was agreed in April 2018.[46]