System-diagnosis of cluster-based parallel architectures

Author

Benkahla, O. ; Aktouf, C. ; Robach, C.

Author_Institution

Lab. de Genie Inf., IMAG, Grenoble, France

fYear

1996

Firstpage

305

Lastpage

309

Abstract

The paper explores the diagnosis of cluster based parallel architectures. A hierarchical strategy which is well suited to such architectures is proposed. This strategy avoids a costly full distributed diagnosis of the network by running an adaptive diagnosis algorithm into each cluster and collecting all the test results at the host level. Key results of the paper include realistic fault and architecture models, an adaptive cluster diagnosis algorithm and a global diagnosis strategy of cluster based parallel machines

Keywords

fault diagnosis; parallel architectures; parallel machines; performance evaluation; adaptive cluster diagnosis algorithm; adaptive diagnosis algorithm; cluster based parallel architectures; cluster based parallel machines; global diagnosis strategy; hierarchical strategy; host level; system diagnosis; test results; Automatic testing; Clustering algorithms; Communication system control; Computer architecture; Concurrent computing; Fault diagnosis; Parallel architectures; Parallel machines; Partitioning algorithms; System testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Processing, 1996. PDP '96. Proceedings of the Fourth Euromicro Workshop on

Conference_Location

Braga

Print_ISBN

0-8186-7376-1

Type

conf

DOI

10.1109/EMPDP.1996.500601

Filename

500601