Computing structures and optical interconnect: friends or foes? Jan M. Van Campenhout Department of Electronics and Information Systems, Ghent University Sint-Pietersnieuwstraat 41, B-9000 Ghent, Belgium ABSTRACT Optical interconnect is claimed to have signi¯cant advantages over electrical interconnect due to the superior physical properties of photon propagation over electron propagation. These advantages translate to the well-known fundamen- tally di®erent bandwidth/distance/cross-section relationships of optical interconnects as compared to their electrical counterparts. Despite the validity of these basic results, too naive or straightforward an application of optical interconnects in electrical systems may not bring about the expected performance gains. In fact, the scienti¯c literature contains several examples of proposed applications of optical interconnect where the real, quanti¯able bene¯ts are at best doubtful. Often, these examples are only intended to demonstrate the feasibility of a technological approach, not its great potential at the systems or application level. Typical °aws one encounters are (i) addressing the wrong problem: one tries to introduce optical interconnect at a location where the electrical interconnect is not the limiting factor; (ii) failing to realize the di®erences between interconnect and communications; and (iii) singling out one property of optical interconnect, while disregarding less bene¯cial properties such as latency, area, power dissipation or optical pathway cost. Capitalizing on the intrinsic potential of optical interconnects in electrical systems requires a holistic approach, that should address the real issues in future electronic systems. To assess the true bene¯ts of replacing an electrical interconnect by an optical one, it is imperative to take the systems context in which the interconnect is to be used into account. By means of examples taken from the computing area, we shall illustrate the profound impact of this context. We then propose a possible parameter space that is intended to capture part of the relevant context information at the link level, thus providing a badly required common domain of discourse for systems designers and optical component designers. Keywords: Optical Interconnect, Interconnect Parameter Space, Interconnect Context 1. INTRODUCTION Using optics in information processing has been a long-standing research goal. In successive waves or generations, attempts have been made at replacing electronic data processing and transport by a photonic or optical counterpart. Several authors 1{3 describe early attempts to build optical processing systems. Free-space Fourier-plane optical computers, ultra-high speed correlators, massively parallel intelligent optical pixel plane architectures are just a sample of the approaches that have been studied at great length. Yet, today virtually all processing is still done electronically, using CMOS processors with internal electrical connections. Optical processing, if any, is still limited to niche applications, and there are no signs that this might change in the years to come. In contrast, quite a di®erent story is to be told about telecommunications, the long- haul transport of information. Spurred by the vastly growing demands of wide-area computer networks, and steady evolution towards the integration of communications networks and computer networks, optical-¯ber communications in the Gigabit-per-second range are commonplace. The use of wavelength division multiplexing techniques (WDM) are further pushing the bandwidth of a single glass-¯ber interconnect into the Terabit-per-second range. The main reason why there has been such breakthrough in communications but not in processing is related to the performance of the mainstream (electronic) technology. In the case of processing, the evolution of both CMOS technology and computer architecture have been no less than phenomenal. Moore's empirical Law, predicting a performance doubling every 18 months, has been validated over and over. Except perhaps for the most demanding niche applications, traditional technology has been able to keep up with the requirements of the applications. In fact, Further author information: E-mail: Jan.VanCampenhout@rug.ac.be