The future of communications is massively parallel

Abstract

Over the past decade, high-speed communications technologies have reached severe scalability limits, from short-reach electrical chip-to-chip interconnects to ultra-long-haul subsea optical fiber cables. While these scalability limits have different origins ranging from saturating high-speed electronics bit rates, to systems approaching fiber Shannon capacities, to energy density/distribution limits, there seems to be only a single long-term viable solution that is common to economically overcome all these limits: massively integrated spatial parallelism.