Abstract | ||
---|---|---|
Abstract: Rollback techniques that use message logging and deterministic replay can be used in parallel systems to recover a failed node without involving other nodes. Distributed shared memory (DSM) systems cannot directly apply message-passing logging techniques because they use inherently nondeterministic asynchronous communication. This paper presents new logging schemes that reduce the typically high overhead for logging in DSM. Our algorithm for sequentially consistent systems tracks rather than logs accesses to shared memory. In an extension of this method to lazy release consistency, the per-access overhead of tracking has been completely eliminated. Measurements with parallel applications show a significant reduction in failure-free overhead. |
Year | DOI | Venue |
---|---|---|
1995 | 10.1109/FTCS.1995.466971 | FTCS |
Keywords | Field | DocType |
failure-free overhead,parallel application,rollback recovery,deterministic replay,shared memory,parallel system,use message,lazy release consistency,failed node,per-access overhead,high overhead,reduced overhead,message passing,parallel systems,distributed computing,asynchronous communication,distributed shared memory,distributed processing,sequential consistency,data loggers,concurrent computing,hardware | Asynchronous communication,Nondeterministic algorithm,Shared memory,Computer science,Parallel computing,Login,Distributed shared memory,Rollback,Distributed computing,Logging,Release consistency | Conference |
ISSN | Citations | PageRank |
0731-3071 | 37 | 1.54 |
References | Authors | |
18 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
gurpreet s suri | 1 | 37 | 1.54 |
b jannsens | 2 | 37 | 1.54 |
W. Kent Fuchs | 3 | 1469 | 279.02 |