Optoelectronic neural-network scheduler for packet switches

Citation
Rp. Webb et al., Optoelectronic neural-network scheduler for packet switches, APPL OPTICS, 39(5), 2000, pp. 788-795
Citations number
18
Categorie Soggetti
Apllied Physucs/Condensed Matter/Materiales Science","Optics & Acoustics
Journal title
APPLIED OPTICS
ISSN journal
00036935 → ACNP
Volume
39
Issue
5
Year of publication
2000
Pages
788 - 795
Database
ISI
SICI code
0003-6935(20000210)39:5<788:ONSFPS>2.0.ZU;2-2
Abstract
A novel, to our knowledge, type of packet scheduler that could significantl y outperform current state-of-the-art schedulers is presented. The operatio n and the design of such a scheduler are discussed, and a fully operational experimental implementation is described. The scheduler uses a neural netw ork in a winner-take-all strategy to optimize decisions on the throughput o f both a crossbar and a banyan switching fabric. The problems of high inter connection density are solved by use of a free-space optical interconnect t hat exploits diffractive optical techniques to generate the required interc onnection patterns and weights. (C) 2000 Optical Society of America.