Universal Replicator performs volume consistent, asynchronous replication.
When establishing a replication to a target secondary volume, the storage system copies all the data to the secondary volume. Depending on the volume size, the creation of the replica takes time. As with TrueCopy, the copying can be optimized using DP volumes.
After the replication is created, the storage system maintains the replica on the secondary volume, transferring each write to the secondary volume. In asynchronous mode, the storage system signals each write completion as soon as it is performed on the primary volume, it then transfers it to the secondary volume (copy after write). Journalling ensures the write order is completely guaranteed so that the secondary volume is crash-recoverable at any point in time.
In the replicating state, no read/write operations are permitted to the secondary volume. To access to the secondary volume, the replication must be paused (placed in the suspended state). Universal Replicator maintains a journal of the primary volume for later pair synchronization. In contrast to TrueCopy, long duration pausing can be tolerated by configuring sufficiently large UR journals. When the replication is deleted the storage system releases the UR journals.
The use of Consistency Groups (CTGs) is mandatory for Universal Replicator.