Intel and Facebook's Open Compute Project has seen the development of server rack architecture that disaggregates resources and shares them across the rack, according to an Intel press release.
In the proposed rack architecture, memory, server, networking and storage resources would be separated into discrete modules, Ars Technica reports. In addition, power and cooling will be shared, rather than having separate functions for each server.
The idea behind disaggregation is to enable a "mix-and-match" approach when elements need to be replaced. With critical components separated, each resource can be decoupled from others and upgraded on its own, improving the lifespan of each resource.
Grouping resource types together also lowers expenses by improving flexibility and reliability. Each rack can generate more computing power when unnecessary parts are eliminated, and resources can be better utilised when shared across CPUs.
Because disaggregating components moves them further apart, Intel is working to avoid slowdowns by developing faster interconnects, capable of speeds up to 100Gbps.
According to Intel, the brand's silicon photonics technology is capable of moving large volumes of data at high speeds, and provides a distinct improvement in cost, power efficiency and reliability over electrical copper-based links or more expensive fibre-optic materials.
However, Intel is not the only company developing silicon photonics solutions. IBM announced its own breakthrough using silicon nanophotonics in December 2012, allowing optical components to be integrated with electrical circuits on a single silicon chip.
"We're excited about the flexibility these technologies can bring to hardware and how silicon photonics will enable us to interconnect these resources with less concern about their physical placement," says Frank Frankovsky, chairman of the Open Compute Foundation and VP of hardware design supply chain at Facebook.
"We're confident that developing these technologies in the open and contributing them back to the Open Compute Project will yield an unprecedented pace of innovation, ultimately enabling the entire industry to close the utilisation gap that exists with today's systems designs," Frankovsky concludes.
The first version of the disaggregated rack is expected to be available in 2014.
Share