Disaster Recovery

Disaster Recovery – the switching on of a second remote data center when a first data center stops functioning due to a disaster, such as a drastic fire, flood, explosion, earthquake or other major event. In order for this switchover to take place a secondary data center has to be in existence and ready to be switched on.

This secondary data center should be geographically remote from the first data center so that it is not destroyed by the same event, such as a hurricane or earthquake. It needs to have the same set of servers, storage devices , network switches, etc., and system and application software so that client systems connected to the damaged data center can be redirected to the disaster recovery (DR) site and continue operating. That requires the stored databases, files and objects to be synchronized between the main and secondary data centers.

It also requires that there be a DR plan which is kept updated and regularly tested to ensure that it works. DR sites are thus expensive to set up and operate, and so often only initiated for mission-critical applications and their IT resources. If the main data center runs its mission-critical applications using virtual machines then these VMs could be re-instantiated in a public cloud, with its facilities used as required for the DR site, a virtual DR site and not a physical one. The same would be true where a data center runs entirely on cloud-native, containerized applications, with containers re-instantiated in the cloud.

An alternative arrangement is to have two data centers, physically remote from each other but otherwise identical, act as reciprocal DR sites for each other, with some amount of IT resource capacity held in reserve at each site for that eventuality.