It’s hard to believe that the Open Compute Project was founded just a little over 18 months ago. At our first Summit, in June 2011, we had 200 participants, one official member (Facebook), and exactly three technology contributions (all from Facebook). Today, as nearly 2,000 people converge on Santa Clara to kick off the fourth OCP Summit, we have more than 50 official members and dozens of contributions from a wide variety of technology suppliers and consumers.
In October 2011 we announced the formation of a non-profit foundation to steer the project, and that organization has continued to evolve. In the last six months we’ve hired our first full-time employee, COO Cole Crawford, and added a new Compliance & Interoperability project. We’ve also seen our first international expansion, with the launch of an OCP Asia-Pacific chapter in Tokyo, and some encouraging engagement with the academic and research communities — including a Purdue University competition to design a biodegradable server chassis.
We’ve also seen some tremendous engagement in the broader OCP community in the last six months. Some examples include:
- New members: More than a dozen new organizations have joined, including storage companies like EMC, Fusion-io, Hitachi, and Sandisk; microprocessor organizations like Applied Micro, ARM, Calxeda, and Tilera; and communications companies like NTT Data and Orange.
- Adoption of open source hardware: More companies are being inspired to think differently about how they build out their infrastructures. Examples include Riot Games, which has decided to purchase systems based on OCP designs and built by Hyve Solutions, and Rackspace, which has decided to design and build its own infrastructure using OCP designs as a starting point.
- Momentum around Open Vault and Open Rack: We launched these two projects at the third OCP Summit in May 2012, and we’re already seeing companies iterating on them in interesting ways and developing creative new technologies to complement them. Examples include new Open Rack designs from Avnet, Delta, Emerson, and Sanmina (many of which are being demonstrated at today’s Summit) and new server designs for the SAS expander slot on the Open Vault spec from Calxeda and Intel.
- Products being delivered: Also at the May 2012 Summit, AMD and Intel announced that they’d been collaborating with Fidelity and Goldman Sachs to develop new boards to suit those companies’ specific workloads — codenamed Roadrunner and Decathlete, respectively. Just seven months later, both companies are here today demonstrating finished products.
- New technology contributions: We’re starting to see significant momentum building behind technology contributions to the project. Some of the new contributions include designs for Fusion-io’s new 3.2TB ioScale card; the designs Rackspace has created for its new infrastructure; and designs for a new storage box that Hyve Solutions built for Open Rack. Facebook is also making a few new contributions, including modifications to Open Rack and Open Vault for use in cold-storage environments and designs for a new, all-flash database server (codenamed Dragonstone) and the latest version of its web server (codenamed Winterfell).
But most exciting of all are a series of new developments that will enable us to take some big steps forward toward better utilization of these technologies. One of the challenges we face as an industry is that much of the hardware we build and consume is highly monolithic — our processors are inextricably linked to our motherboards, which are in turn linked to specific networking technology, and so on. This leads to poorly configured systems that can’t keep up with rapidly evolving software and waste lots of energy and material.
To fix this, we need to break up some of these monolithic designs — to disaggregate some of the components of these technologies from each other so we can build systems that truly fit the workloads they run and whose components can be replaced or updated independently of each other. Several members of the Open Compute Project have come together today to take the first steps toward this kind of disaggregation:
- Silicon photonics: Intel is contributing designs for its forthcoming silicon photonics technology, which will enable 100 Gbps interconnects — enough bandwidth to serve multiple processor generations. This technology also has such low latency that we can take components that previously needed to be bound to the same motherboard and begin to spread them out within a rack.
- “Group Hug” board: Facebook is contributing a new common slot architecture specification for motherboards. This specification — which we’ve nicknamed “Group Hug” — can be used to produce boards that are completely vendor-neutral and will last through multiple processor generations. The specification uses a simple PCIe x8 connector to link the SOCs to the board.
- New SOCs: AMD, Applied Micro, Calxeda, and Intel have all announced support for the Group Hug board, and Applied Micro and Intel have already built mechanical demos of their new designs.
Taken together, these announcements will enable data center operators to build systems that better fit the workloads they need to run and to upgrade through multiple generations of processors without having to replace the motherboards or the in-rack networking. This should in turn enable real gains in utilization and unlock even more efficiencies in data center construction and operations.
There’s a lot of work left to do here, but it’s never been more important for us to take these steps. As our lives become more connected, and as more devices and applications generate more data, we will face compute and storage challenges that existing technologies cannot handle efficiently.
But we have one big advantage, as we face these challenges: We are doing this work together, in the open, and everyone has a chance to contribute — to help ensure that all the technologies we develop and consume are as scalable as possible, as efficient as possible, and as innovative as possible.