New Report Examines Potential Interconnect Technologies & How To Accelerate

The potential advantages of optical interconnects for board-to-board, chip-to-chip, and even on-chip interconnects have been touted by optical researchers for many years, and the imminent demise of copper interconnects has been predicted in some circles for more than a decade. Nevertheless, while optical technology has successfully displaced copper wiring in longer distance telecommunication as well as many enterprise networking applications, to date there has been no commercially viable implementation of optical interconnects for applications inside of a box. Even for very short-distance box-to-box links (<10 m), where optical interconnects have started to have a commercial presence, high-bandwidth copper cabling still predominates. The predicted evolution of supercomputers, high-end servers, routers, and storage systems suggests that the need for extremely high bandwidth-densities might create realistic commercial opportunities for “inside-the-box” optical interconnects within the next decade, but many substantial barriers will need to be overcome. In an effort to identify these barriers, produce recommendations to overcome these barriers, and generate an industry-wide common understanding of where we are today and what the future prospects are for short-distance optical interconnects, a workshop, sponsored by DARPA, was held in Burlingame, CA, on October 21 and 22, 2004. This report includes many inputs from that workshop. The one-and-a-half-day roadmapping workshop explored the questions of if, when, and how optical interconnects can begin to replace copper in box-to-box, board-to-board, chip-to-chip, and eventually on-chip applications. The participants worked on defining a roadmap to take us forward from today’s early box-to-box applications of parallel optics toward commercial implementation of inside-the-box optical backplanes in high-end systems and onward toward widespread use of optics for ever shorter-reach applications, perhaps eventually including on-chip interconnects. The goals were to identify the key barriers that must be overcome at each key distance before optics can displace copper and to generate a set of recommendations on approaches to overcoming these barriers. ... http://tinyurl.com/2sfxcc

Note: http://www.researchandm... http://tinyurl.com/2sfxcc