A transaction that invokes operations at several servers. A transaction that invokes operations at several servers

A transaction that invokes operations at several servers.

• A transaction that invokes operations at several servers.

In a distributed environment, a coordinator is needed

• In a distributed environment, a coordinator is needed

• Client sends an openTransaction to the coordinator

• Other servers that manage the objects accessed by the transaction become participants.
• The coordinator provides a join method interface

Atomicity principle requires that either all the distributed operations of a transaction complete, or all abort.

• Atomicity principle requires that either all the distributed operations of a transaction complete, or all abort.

• At some stage, client executes closeTransaction(). Now, atomicity requires that either all participants (remember these are on the server side) and the coordinator commit or all abort.

• What problem statement is this?

Consensus, but the system is asynchronous!!

• Consensus, but the system is asynchronous!!

• So, need to ensure safety property in real-life implementation. Never have some agreeing to commit, and others agreeing to abort.

• First cut: one-phase commit protocol. The coordinator unilaterally communicates either commit or abort, to all participants (servers) until all acknowledge.

• Doesn’t work when a participant crashes before receiving this message (partial transaction results that were in memory are lost).
• Does not allow participant to abort the transaction, e.g., under error conditions.

Consensus, but it’s impossible in asynchronous networks!

• Consensus, but it’s impossible in asynchronous networks!

• So, need to ensure safety property in real-life implementation. Never have some committing while others abort. Err on the side of safety.

• Alternative: Two-phase commit protocol

• First phase involves coordinator collecting a vote (commit or abort) from each participant
• Participant stores partial results in permanent storage before voting
• Now coordinator makes a decision
• if all participants want to commit and no one has crashed, coordinator multicasts “commit” message
• Everyone commits
• If participant fails, then on recovery, can get commit msg from coord
• else if any participant has crashed or aborted, coordinator multicasts “abort” message to all participants
• Everyone aborts