Architecture Alternatives for Single System Image Clusters


Abstract
The availability of high-speed networks and increasingly powerful commodity microprocessors are making the usage of clusters, or networks, of computers an appealing vehicle for cost effective parallel/dependable computing. Clusters, built using commodity-of-the-shelf (COTS) hardware components as well as free, or commonly used, software, are playing a major role in redefining the concept of supercomputing. In this talk, we focus on the issues involved in making loosely coupled networked computers transparently appear and work as a single system. This properly of cluster is popularly called as "single system image", SSI in short.

A single system image properly of clusters can be created either by software or/and hardware. In this talk, we focus on techniques for achieving a single system image by software means-cluster underware (i.e, at OS kernel level), cluster middleware (i.e., a layer between applications and operation systems), and application level. We also discuss the use of these techniques in research and commercial software systems (such as MOSIX, UnixWare, PARMON, Nimrod, etc.) supporting a single system image. We bring out merits and demerits of these SSI techniques and conclude the talk with highlighting possible directions that cluster computing research and market is heading towards.

Click here to download the Single System Image Paper (which is in Powerpoint format).

Appeared in: Proceedings of the 1999 Conference on High Performance Computing on Hewlett-Packard Systems (HiPer'99), Tromse, Norway, 1999.