Dealing with failures. In the description of the algorithm we assumed no failures. Let us now study what happens if failures occur. According to our objectives, transactions should commit, even if some of the participants abort, and most importantly, sites should not become blocked. We consider three possible types of failure: site crash, message loss, and network partition.
Appears in 2 contracts
Sources: Independent Updates and Incremental Agreement in Replicated Databases, Independent Updates and Incremental Agreement in Replicated Databases
Dealing with failures. In the description of the algorithm we assumed no failuresfail- ures. Let us now study what happens if failures occur. According to our objectives, transactions should commitcom- mit, even if some of the participants abort, and most importantly, sites should not become blocked. We consider three possible types of failure: site crash, message loss, and network partition.
Appears in 1 contract
Sources: Independent Updates and Incremental Agreement in Replicated Databases