Geek Answers Handbook

Articles about replication schemes / algorithms?

I am developing a distributed system with a specific data stream in it. I would like to guarantee that at least N nodes have almost current data at any given time. I do not need full consistency, only final consistency (ti for any moment in time, the current snapshot of the data should eventually appear at least on N nodes. It is difficult to define the term “current” here, but still). Nodes can fail and return at any time, and there is no single "central" node.

O overflow! Call me good docs that describe replication schemes. I have still found one: Consistency management in optimistic replication algorithms and a wider and more recent article by the same author: Optimistic replication .

+5
source share
2 answers

Most of this trick is finding your exact requirements, and yours still sound pretty vague. Do you just need to support such operations?

  • Update key K to V.
  • Look at the somewhat recent value of key K.

, . , , . , , ? , , , , ? ?

? netsplit -, . , ?

- , ...

  • K?
  • K V ', V?

, / ? / ? , -- ; / , .

, . .

  • - . Replicas ( NTP) . . , //, . (, , .) , . , Bigtable. , , " - ", , , .
  • Paxos (., , " - Single Sign-On" , , , , " , 1 " " ". , , " ", , , . , , . , , , , , , Paxos, . , ...
+1

RDBMS , SQL Server 2008 ( 2005 )

0

More articles:


All Articles