Can you outline some specific ways remote data replication is used for backup and disaster recovery? For example, how does CDP fit into the picture?
Replication is an extension of mirroring, and it's for protecting data. Some people might be using a form of replication already, without realizing it. For example, if you are mirroring two disks, or two storage systems, or even two servers, that is a form of replication.
But replication is typically associated with some level of distance -- across the room, across the campus, across the city or even across the continent. What it's really there for is to ensure that you have two copies of data (or three, or however many replicas you need) for availability or disaster recovery purposes.
A couple years ago, CDP, or continuous data protection, was a big buzzword. As its name implies, it is about making sure that data is continuously copied between two systems. But we need to talk about another aspect -- point-in-time copies or snapshots.
When you are constantly replicating data remotely, if it corrupts at one side, it corrupts on the other side as well. So, this is why it's important to have that point-in-time copy for you to roll back to when the data is good and intact. CDP gives you that ability to go back to a dynamic point in time and say, "That's where I want to recover from." Near CDP relies on snapshots. Snapshots are typically set up at some predetermined interval, whether it's hourly, every 15 minutes, daily -- whatever it happens to be. So, the intervals are longer than with true CDP.
This was first published in March 2013