Optical Interconnection Networks

Abstract

Publisher Summary This chapter discusses the optical interconnection networks. Ongoing advances in high-speed integrated circuit device technology along with increasing demand in data communication, digital signal and image processing, neural network, and machine vision systems necessitate the design of efficient interconnections topology. In electrical interconnections, the scope of the systems is limited by interference and the planar nature of very large-scale integrated (VLSI) circuits. The chip speed—in particular, is limited by interconnect delays. Intrachip connections pose further difficulty. Optical interconnects, because of their inherent parallelism, high speed, and negligible mutual interference, offer a reasonable solution to this intricate problem. Optical techniques may provide an alternative means for fast, secure interconnections of the devices directly to the interior of a chip. Optical computing research involving optical bistable devices, nonlinear Fabry–Perots, hybrid electrooptic devices, and optical interconnects has contributed to two distinct, parallel interconnect architectures. One approach uses integrated optics to interconnect optical logic devices, while the other involves 2D arrays of devices interconnected in free-space. Free-space interconnection, in turn, is classified as either focused (obtained by lens, beamsplitter, etc.) or unfocused (obtained by imaging light through holographic elements). The index-guided waveguide interconnects may be either space variant or space invariant.