Synchronous mirroring utilizing storage based replication

This topic describes synchronous mirroring utilizing storage-based replication.

Synchronous replication in the storage layer continuously updates a secondary (target) copy of a disk volume to match changes made to a primary (source) volume. A pair of volumes are configured in a replication relationship, during which all write operations performed on the source are synchronously mirrored to the target device.

The synchronous replication protocol guarantees that the secondary copy is constantly up-to-date by ensuring that the primary copy is written only if the primary storage subsystem received an acknowledgment that the secondary copy has been written. The paired volumes typically reside on two distinct and geographically separated storage systems communicating over a SAN or LAN link.

Most synchronous replication solutions provide a capability to perform an incremental resynchronization of data when switching between primary and secondary storage systems. After the failure of the primary volume (or the failure of the entire storage subsystem or site), users perform a failover, which suspends the relationship between the given pair of volumes and turns the target volume into a primary. When a volume enters the suspended state, a modification bitmap is established to keep track of the write operations performed on that volume to allow for an efficient resynchronization.

Once the operation of the original primary volume has been restored, a failback is executed to resynchronize the content of the two volumes. The original source volume is switched to the target mode, after which all modified data tracks (those recorded in the modification bitmap) are copied from the original target disk. The volume pair can then be suspended again and a similar process performed to reverse the volumes' roles, thus bringing the pair into its initial state.