This preview shows page 1. Sign up to view the full content.
Unformatted text preview: e provide the new data to a refetch, but the server is sharded (callback state spread across two or more machines)? Then to keep the clients seeing the updates in order, either the updates need to be (somehow) ordered across shards, or the refetch must wait until all invalidates have been applied. (Optimization: the server could broadcast the update, but to be serializable, it needs to ensure that everyone sees the updates in the same order.) Can we be more efficient than write-‐through? Write-‐through means that we must contact the server on every file modification. Imagine the pesky little brother saying: there I changed the file. There I changed it again. And again. Pretty soon you’d say: enough already! Just tell me when you log out! Write back cache coherence allows for changes to be kept at the client. (Of course, this means that if the client crashes in the meantime, you might lose some of its updates.) Illustrate state machine for write back cache coherence: owned, read...
View Full Document
This document was uploaded on 04/04/2014.
- Spring '14