An argument in favor of the presumed commit protocol

The authors argue in favor of the presumed commit protocol by proposing two new presumed commit variants that significantly reduce the cost of logging activities associated with the original presumed commit protocol. Furthermore, for read-only transactions, they apply their unsolicited update-vote optimization and show that the cost associated with this type of transactions is the same in both presumed commit and presumed abort protocols, thus, nullifying the basis for the argument that favors the presumed abort protocol. This is especially important for modern distributed environments which are characterized by high reliability and high probability of transactions being committed rather than aborted