Talk:Consensus (computer science)

From Wikipedia, the free encyclopedia

The paper "The consensus problem in fault-tolerant computing" from Michael Barborak, Miroslaw Malek and Anton Dahbura (1993) is a giant survey about the topic (25 years of research), and could greatly improve this article.

It would also be helpful to have a decent explanation of the intent of consensus. All of the papers on the internet are pay-per-download. However, the information about how the algorithms work are everywhere, and none of them describe the use of a consensus algorithm. It is understandable that distributed transactions must reach agreement, and through consensus can decide. But no one mentions what the algorithm is agreeing upon. Is it states of the replicates, initial conditions given to replicas, the end result of a computation (i.e. everyone must agree on what has been done), agreement upon where to send the data or what to do with it, or is it only an agreement to ensure that no one has failed during some time?? —Preceding unsigned comment added by Jon.gallagher.04 (talk • contribs) 18:26, 14 April 2008 (UTC)