Given a distributed causally-consistent database (a replicated key-value store where writes are propagated in causal order, in the style of this paper), what are common techniques for "scaling" the number of replicas up or down?

In particular, when a new replica joins the DB, how do we efficiently bring it "up to speed" relative to other replicas?

I'm looking for references to papers or existing industry implementations.


This site is temporarily in read only mode and not accepting new answers.

Browse other questions tagged .