Today, let's sing in praise of replication. Replication has been the mainstay technology of disaster recovery strategies for quite some time now. Data is copied to a remote site, where it is archived and in an emergency, it is brought back online again at a primary site.
Rules regarding remote sites are few, but important. Rule one is usually that the more distance between the sites, the better. Rule two is that the remote site ought to be in a geologically stable area. Rule three says that recovery ought to be as fast as possible.
The first rule works to isolate events at the two sites. Rule two speaks well for property values in North and South Dakota. And rule three is a good indicator of why Iron Mountain is evolving from being a simple tape depository.
Replicated data has value far beyond DR copies at a remote site, however. Consider some additional usesfor replicated volumes if the replicas are stored locally. The following list is brief, but serves to show where this can take you.
Replicate a volume to local cheap storage, and then perform high-workload services from the replica. Under such a scenario a database might be copied from Fibre Channel devices to SATA, with the following results: the worry about overstepping maintenance windows is eliminated because of the speed of disk-to-disk copying; reports and backups to tape - cycle-suckers in any environment - can be written from the replica; and performance on the primary system remains unaffected.
A further benefit: while the replica remains on the near line device, recoveries will always be faster than they could ever be by tape.
Replication scenario two: replicate a data set and ship the replica off to your R&D team. You have now provided them with a data set for their test beds that exactly reflects the data in the real world. Looking for payback? All the time that the quality assurance group no longer invests in building and rebuilding test bed data may more than pay for your investment in an efficient replication product, and the test environment itself will be far more robust.
Replication scenario three: replicate a data set and let your IT support team use it as they prepare to roll out the next version of some key software they are supporting. They can test the new release against a real-world environment, and by the time the upgrade goes live you can be sure that the number of "gotchas" in the upgrade will be far fewer than what you are used to seeing.
Replication's value - particularly when it is automated and doesn't take a lot of time out of an admin's day - delivers value far beyond what most IT managers expect to get from it.
Think about how you might be able to use it, and do some research on the topic.