Demonstration of a compact low-power 250-Gb/s parallel-WDM optical interconnect
ABSTRACT In this letter, we demonstrate error-free operation of a 12-fiber ×4-wavelength ×5.21-Gb/s parallel-wavelength-division-multiplexed (PWDM) optical link. The 250-Gb/s transmitter and receiver assemblies each have a 5×8-mm footprint and consume a combined power of 1.5 W. To our knowledge, this is the first publication of a fully functional PWDM optical interconnect as well as the highest demonstrated bandwidth per unit area and bandwidth per unit power consumption for any multiple-channel fiber-optic interconnect. This technology is intended for short-distance high-bandwidth-density applications such as multiprocessor computer backplanes.
- SourceAvailable from: cs.fsu.edu[Show abstract] [Hide abstract]
ABSTRACT: Optical interconnect architectures with electronic buffers have been proposed as a promising candidate for future high speed interconnections. Out of these architectures, the OpCut switch achieves low latency and minimizes optical-electronic-optical (O/E/O) conversions by allowing packets to cut-through the switch whenever possible. In an OpCut switch, a packet is converted and sent to the recirculating electronic buffers only if it cannot be directly routed to the switch output. In this paper, we study packet scheduling in the OpCut switch, aiming to achieve overall low packet latency while maintaining packet order. We first decompose the scheduling problem into three modules and present a basic scheduler with satisfactory performance. To relax the time constraint on computing a schedule and improve system throughput, we further propose a mechanism to pipeline packet scheduling in the OpCut switch by distributing the scheduling task to multiple “sub-schedulers.” An adaptive pipelining scheme is also proposed to minimize the extra delay introduced by pipelining. Our simulation results show that the OpCut switch with the proposed scheduling algorithms achieve close performance to the ideal output-queued (OQ) switch in terms of packet latency, and that the pipelined mechanism is effective in reducing scheduler complexity and improving throughput.Journal of Lightwave Technology 01/2012; 30(12):1869-1881. · 2.86 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Nowadays, multiprocessor systems are reaching their limits due to the large interconnection bottleneck between chips, but recent advances in the development of optical interconnect technologies can allow the use of low cost, scalable and reconfigurable networks to resolve the problem. In this paper, we make an initial evaluation of the performance gain on general network reconfigurability. In a next stage, we propose an optical system concept and describe a passive optical broadcasting component to be used as the key element in a broadcast-and-select reconfigurable network. We also discuss the available opto-electronic components and the restrictions they impose on network performance. Through detailed simulations of benchmark executions, we show that the proposed system architecture can provide a significant speedup for shared-memory machines, even when taking into account the limitations imposed by the opto-electronics and the presented optical broadcast component.
Article: Dielectric Multilayer FiltersWavelength Filters in Fibre Optics. 01/2006;