Dr. Darrell Long

Reliability of Replicated Data Objects

Proceedings of the International Performance Conference on Computers and Communication

March 1989View full paper

Abstract

Improved fault tolerance of many applications can be achieved by replicating data at several sites. This data redundancy requires a protocol to maintain the consistency of the data object in the prescence of site failures. The most commonly used scheme is voting. Voting and its variants are unaffected by network partitions. When network partitions cannot occur, better performance can be achieved with availability copy protocols. Common measures of dependability include reliability, which is the probability that a replicated object will remian constantly available over a fixed time period. We investigate the reliability of replicated data objects managed by voting, available copy and their variants. Where possible, closed-form expressions for the reliability of the various consistency protocols are derived using standard Markovian assumptions. In other cases, numerical solutions are found and validated with simulation results.

BibTeX

@inproceedings{long-ipccc89,
  author       = {Darrell D. E. Long and Jehan-Fran\c{c}ois P\^aris and John L. Carroll},
  title        = {Reliability of Replicated Data Objects},
  booktitle    = {Proceedings of the International Performance Conference on Computers and Communication},
  publisher    = {IEEE},
  month        = mar,
  year         = {1989}
}
View full paper