METASYSTEMS - AN APPROACH COMBINING PARALLEL-PROCESSING AND HETEROGENEOUS DISTRIBUTED COMPUTING SYSTEMS

Citation
As. Grimshaw et al., METASYSTEMS - AN APPROACH COMBINING PARALLEL-PROCESSING AND HETEROGENEOUS DISTRIBUTED COMPUTING SYSTEMS, Journal of parallel and distributed computing, 21(3), 1994, pp. 257-270
Citations number
31
Categorie Soggetti
Computer Sciences","Computer Science Theory & Methods
ISSN journal
07437315
Volume
21
Issue
3
Year of publication
1994
Pages
257 - 270
Database
ISI
SICI code
0743-7315(1994)21:3<257:M-AACP>2.0.ZU;2-T
Abstract
A metasystem is a single computing resource composed of a heterogeneou s group of autonomous computers linked together by a network. The inte rconnection network needed to construct large metasystems will soon be in place. To fully exploit these new systems, software that is easy t o use, supports large degrees of parallelism, and hides the complexity of the underlying physical architecture must be developed. In this pa per we describe our metasystem vision, our approach to constructing a metasystem testbed, and early experimental results. Our approach combi nes features from earlier work on both parallel processing systems and heterogeneous distributed computing systems. Using the testbed, we ha ve found that data coercion costs are not a serious obstacle to high p erformance, but that load imbalance induced by differing processor cap abilities can limit performance. We then present a mechanism to overco me load imbalance that utilizes user-provided callbacks. (C) 1994 Acad emic Press, Inc.