Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Modular Opto-Mechanical Design of Free-Space Optical Interconnect System for Massively Parallel Processing

Not Accessible

Your library or personal account may give you access

Abstract

Free-space optical interconnections can provide parallel access for many communication patterns with a large bandwidth[1, 2]. The challenge in the realization of such systems relies in finding good solution for the alignment and packaging of the electro-optical components. We present a new modular building block that will address these challenges in a free-space optical network for massively parallel processing (MPP). Our component and system ideas offer a simple method to fabricate and build the optical network assuring reliable and convenient alignment and packaging. We use micro-lens arrays which are pre-aligned and packaged as modular building blocks to relay optical beams from one module to the next. We can apply our modular building block ideas to our previous system of a 64-channel free-space optical interconnection[1] and its wavelength division multiplexing (WDM) extensions[3] to realize a more reliable and easier to build system. In the current network architecture, as well as in the previous ones, we use a nonblocking and reconfigurable interconnection cached network (ICN)[4].

© 1997 Optical Society of America

PDF Article
More Like This
Massive Optical Interconnections (MOI): Interconnections for Massively Parallel Processing Systems

S. Araki, M. Kajita, K. Kasahara, K. Kubota, K. Kurihara, I. Redmond, E. Schenfeld, and T. Suzaki
OMA3 Optical Computing (IP) 1995

A Comparison of Massively Parallel Interconnect Topologies Employing Optical Highways

Julian A. B. Dines, David T. Neilson, John F. Snowdon, and Brian S. Wherrett
OThD.1 Optics in Computing (IP) 1997

System Analysis of Global Free-Space Optical Interconnect

V. Morozov, J. Neff, and H. J. Zhou
OThD.13 Optics in Computing (IP) 1997

Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.