One-Pagers‎ > ‎

OP10: Eventual consistency

Finish the "recipe" for eventual consistency in multi-master replication what was started during the debate session. The goal is to reach at least the same level of precision as that used in Lampson's "redo logging" example.

Please keep in mind that the recipe ought to accommodate more than just one system (the analogy we used in class is that your problem definition represents a grammar, and the various system designs are strings accepted or rejected by the grammar; the definition should therefore be sufficiently precise to serve as a grammar, and broad enough to be useful).