CSC 7103 20090917

CSC 7103 20090917 - CSC 7103 9/17/2009 Page 1 of 2 RPC The...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
CSC 7103 9/17/2009 Page 1 of 2 RPC The most important failures are client and server failures. Message failures should be handled by the reliable transport functionality. When a failure occurs in the client or server we must make sure that the requested operation is not executed multiple times. Idempotent operations can be executed many times and the result will not change. Ideally all processes are idempotent, but they are not. RPC semantics: At-most-once semantics – executed zero or one times, no more. “maybe” semantics – may be executed many times may not Exactly once – executed exactly once How do we ensure exactly once semantics? Client crashing How do we handle orphan calls that may spawn other calls? Server can kill the orphans Timeout on responses (see first item) The client could remember all transactions and pick back up when it comes back online (bad idea) Transactions Transactions are the highest level representation of communication between processes. Synchronization
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 2

CSC 7103 20090917 - CSC 7103 9/17/2009 Page 1 of 2 RPC The...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online