Creating a Remote Replication Pair
HyperReplication copies data from a local storage system to a remote storage system for remote DR. HyperReplication includes synchronous remote replication and asynchronous remote replication.
Context
HyperReplication is a data mirroring technology that copies data on a local storage device to a remote storage device. This technology allows multiple data copies to be maintained at two or more sites, eliminating data loss risks. HyperReplication covers HyperReplication/S (synchronous remote replication) and HyperReplication/A (asynchronous remote replication).
- In HyperReplication/S mode, when data is written to a primary resource, the write request is also sent to a secondary resource. After both the primary and secondary resources return a write success response, the storage system sends a write success message to the host. In this way, data on the primary resource is consistent with that on the secondary resource, preventing data loss in the event of a disaster.
- Each data write request is successful only after being responded and executed simultaneously on the local and remote storage devices. If the local and remote storage devices are far away from each other, the write latency of foreground applications is high, adversely affecting normal service running. Therefore, HyperReplication/S is applicable to disaster recovery scenarios where the local storage device is close to the remote storage device, for example, same-city disaster recovery.
- HyperReplication/S ensures data consistency between the primary and secondary resources. If the local storage device is faulty, data on the remote storage device is consistent with that on the local storage device, and no data is lost. Therefore, HyperReplication/S is mainly used in disaster recovery scenarios that have strict requirements on data security.
- In HyperReplication/A mode, when data is written to a primary resource, the primary resource records the data changes. After the primary resource returns a write success message, the storage system informs the host that the data is written successfully. When the amount of data changes reaches a certain value, all the data changes are updated to a secondary resource on the remote storage device at a time. The write latency of foreground applications is remarkably reduced so that the system throughput and performance are improved.
- The write latency of foreground applications is independent of the distance between the local and remote storage devices. Therefore, HyperReplication/A is applicable to disaster recovery scenarios where the distance between the local and remote storage devices is long or the network bandwidth is low.
- During asynchronous remote replication, if data on the local storage device is not synchronized to the remote storage device and the local storage device becomes faulty, the data that has not been synchronized will be lost. Therefore, HyperReplication/A is mainly used in disaster recovery scenarios that have higher requirements on system response time.
- Creating an Asynchronous Remote Replication Pair
This section describes how to create a remote replication pair to periodically synchronize data from the primary site to the secondary site. This is applicable to scenarios that the two sites are over long distance or the bandwidth is limited. - Creating a Synchronous Remote Replication Pair