Free dimensions-an effective approach to achieving fault tolerance in hypercube

Author

Raghavendra, C.S. ; Yang, P.-J. ; Tien, S.-B.

Author_Institution

Washington State Univ., Pullman, WA, USA

fYear

1992

fDate

8-10 July 1992

Firstpage

170

Lastpage

177

Abstract

In the n-dimensional hypercube, Q/sub n/, for large n, faults can occur with relatively high probability. How to use the inherent redundancy present in the hypercube to obtain fault tolerance is discussed, along with computing in faulty hypercubes. The authors study the fault tolerance independently present in hypercubes by defining and using the concept of free dimensions. Briefly, in Q/sub n/, a dimension is said to be free if no pair of nodes across the dimension link are both faulty. Efficient algorithms are presented for finding free dimensions, given a set of faulty nodes, and it is shown that at least n-f+1 free dimensions exist with f>

Keywords

fault tolerant computing; hypercube networks; parallel architectures; broadcasting algorithms; embedding; emulation; fault tolerance; free dimensions; hypercube; inherent redundancy; n-dimensional hypercube; reconfiguration; routing; Emulation; Fault tolerance; Fault tolerant systems; High performance computing; Hypercubes; Intelligent networks; Parallel processing; Partitioning algorithms; Redundancy; Robustness;

fLanguage

English

Publisher

ieee

Conference_Titel

Fault-Tolerant Computing, 1992. FTCS-22. Digest of Papers., Twenty-Second International Symposium on

Conference_Location

Boston, MA, USA

Print_ISBN

0-8186-2875-8

Type

conf

DOI

10.1109/FTCS.1992.243603

Filename

243603