Global Copy for ESS is a Business Continuity and Disaster Recovery tool that supports disk mirroring over great distances.
Global Copy for ESS was previously known as PPRC-eXtended Distance (PPRC-XD). This is a microcode feature of the Enterprise Storage Server that supports asynchronous data replication (disk mirroring). The asynchronous nature of the replication process allows for much greater distances between the primary and secondary storage subsystems than with synchronous PPRC (Metro Mirror for ESS).
There is no performance penalty for distance between the storage systems. Because of this, there is no increase in the I/O response time as there is with synchronous PPRC. Instead, production writes occur much as they do in a local-only environment: Once the data has been written to the local DASD subsystem, control is returned to the application running at the local site without any additional delay while the data is being mirrored to the remote site.
The hardware configuration shown here is identical to what might be used in a Metro Mirror or synchronous PPRC implementation, but because of the asynchronous nature of Global Copy for ESS, much greater distances can be achieved than with synchronous PPRC.
The local (or primary) storage subsystem buffers the changed data in cache until it can be successfully written to the remote (or secondary) DASD subsystem. In this way, the data is mirrored off-site without any performance penalty. However there is nothing automatic in this type of implementation that provides data consistency. This is because without additional intervention or automation, this facility does not by itself provide data consistency. In fact, the secondary data may be is considered “fuzzy” or unusable, as the secondary writes occur without regard to the order that the primary writes have occurred.
Therefore, additional process must be implemented in order to ensure that the secondary copy of the data is actually useable for recovery. In general, this can occur by either halting processing at the production site (allowing all changed data to be written) or otherwise by temporarily terminating the mirroring environment.
Global Copy for ESS (PPRC-XD) is frequently used when migrating data from one disk subsystem to another, or sometimes for Disaster Recovery when a "fuzzy" copy of the data may be acceptable.
The logical diagram at the right helps to illustrate how point-in-time consistency can be achieved in this configuration: Most of the time, the data from the primary volumes are being mirrored asynchronously to the secondary volumes. This avoids any impact to the production applications but unfortunately, the secondary copy is fuzzy and may be of limited value for recovery.
This problem can be addressed operationally by periodically changing to synchronous mirroring at times when the production applications can tolerate the additional response time delay induced by the synchronous write. Once the secondary volumes have been fully synchronized, the secondary volumes can be FlashCopied to a tertiary set of volumes. This tertiary set of volumes provides a point-in-time and fully time consistent set of data for Disaster Recovery.
Once the FlashCopy commands have been issued, the data mirroring can be changed back to asynchronous to avoid any impact to the applications at the primary site.
This document was printed from http://recoveryspecialties.com/