This preview shows page 1. Sign up to view the full content.
Unformatted text preview: use everyone’s TTL will have expired.] [Another anecdote: NFS server. If it crashes, what does it need to do when it reboots? Nothing! Clients can simply retry any RPC’s they had in progress, since there is no callback state at the server.] You can think of NFS as the pesky little sister repeatedly asking: has this item changed? Has this item changed? Until you want to scream, I’ll tell you when it does! Reached here Review: Table of write-‐through, write-‐back, TTL, callback TTL can be inefficient, since if data is being updated frequently, need to check server repeatedly even if data hasn’t changed. So: callbacks. Record state at server as to who has which cached copy, so that server can tell client when its cached data is invalid. Illustrate state machine for write-‐through cache coherence. State is per-‐client/per-‐
memory object. A memory location can be read-‐only or invalid on each client. To keep things consistent, when a node updates the server (write-‐through), serve...
View Full Document
This document was uploaded on 04/04/2014.
- Spring '14