Data center replication - a key step in disaster recovery
The Atlantic Hurricane season has officially started which means that most organizations would have made plans of some kind of disaster recovery for their data centers. This usually involves developing failover centers and backing up data in the event that the primary data center goes down. And this is where data center replication comes in.
As I mentioned in my last blog post, replication of a data center is a preferred option in scenarios where the target data centers will be equipped with identical hardware and OS as that of the source data center - resulting in a complete replication of the environment itself.
In tech-speak, we define replication as a method in which in various infrastructure computing devices (both physical and virtual) share information across data centers so as to ensure consistency between resources to improve reliability and fault-tolerance.
As with any other kind of migration, replication of a data center involves a set of key considerations primarily around the volume and type of data to be replicated and the number of sites to be migrated:
1- Size/volume to data to be replicated, velocity/Speed require replicating data and variety/type of data.
2- Distance from the source datacenter, number of sites/ domains and security policies to be migrated.
The migration itself may be done in several steps. For instance, a data center may be replicated with the same physical and virtual environment and the same storage or the storage may get replicated separately. In either case, the migration process would require proper planning with focus on bundling applications for relocation and managing licenses for both physical and/or virtual environments. There is a lot of potential to reduce both capex and opex costs at this stage. One sure shot way would be to use virtualization to reduce the count of servers and thereby reduce the associated license costs.
The importance of being prepared with DR cannot be stressed enough. A quick check of the news headlines including the Amazon outage in 2011 and the more recent outage of the French government's accounts payable system - Chorus, makes it clear that system downtime can be crippling for any organization. Replication of a data center to create a new DR site offers organizations a plausible approach towards balancing risks with business needs.