locks, but non-conflicting ones benefit from parallelized, multi-threaded execution in multi-core environments.
In an HSB database, transaction logs are sent to the Secondary server by way of a replication protocol. In
order to preserve the database consistency in the presence of fail over, the replication protocol is built very much
on the same principles as physical log writing: the transaction order is preserved, and commit records denote
committed transactions. If a fail over happens, the standby server performs a similar database recovery as if
a transaction log was used: the uncommitted transactions are removed and the committed ones are queued for
execution.