Abstract

A novel, to our knowledge, type of packet scheduler that could significantly outperform current state-of-the-art schedulers is presented. The operation and the design of such a scheduler are discussed, and a fully operational experimental implementation is described. The scheduler uses a neural network in a winner-take-all strategy to optimize decisions on the throughput of both a crossbar and a banyan switching fabric. The problems of high interconnection density are solved by use of a free-space optical interconnect that exploits diffractive optical techniques to generate the required interconnection patterns and weights.

© 2000 Optical Society of America

Programmable optoelectronic neural network for optimization

Keith J. Symington, Yves Randle, Andrew J. Waddie, Mohammed R. Taghizadeh, and John F. Snowdon
Appl. Opt. 43(4) 866-876 (2004)

FPGA-based implementation of two-step schedulers for modular optical interconnection networks

Justine Cris Borromeo, Isabella Cerutti, Piero Castoldi, Rosula Reyes, and Nicola Andriolli
J. Opt. Commun. Netw. 13(5) 116-125 (2021)

Scheduling with Machine-Learning-Based Flow Detection for Packet-Switched Optical Data Center Networks

Lin Wang, Xinbo Wang, Massimo Tornatore, Kwang Joon Kim, Sun Me Kim, Dae-Ub Kim, Kyeong-Eun Han, and Biswanath Mukherjee
J. Opt. Commun. Netw. 10(4) 365-375 (2018)

Pseudo-Banyan Optical WDM Packet Switching System With Near-Optimal Packet Scheduling

Maria C. Yuang, Po-Lung Tien, and Shih-Hsuan Lin
J. Opt. Commun. Netw. 1(3) B1-B14 (2009)

Self-routing crossbar packet switch employing free-space optics for chip-to-chip interconnections

Thomas J. Cloonan and Anthony L. Lentine
Appl. Opt. 30(26) 3721-3733 (1991)