In Part I, we discussed the importance of data loads, de-duplication, encryption, and hardware/software configuration, as best practices for Cloud backup and recovery.

The “data on demand” environment of modern day computing underscores the need for putting in place effective methods of data recovery.  IT administrators, charged with the task of ensuring recoverability of data at all times, must be aware of world best practices in this area and take the steps to institute these practices within the organization.

Planning for recovery: Data recovery from Cloud based backup systems happens in all or any of the following situations:

  1. Accidental deletion of a file or folder
  2. System crashes due to hard disk failures and other related failures
  3. Natural disasters that wipe out a whole data center
  4. Creation of a new hot site or data center for failover purposes
  5. Setting up a data center for distributed computing

The recovery demands may include strict pre-defined Recovery Time Objectives (RTO) and Recovery Point Objectives (RPO), cascading into definitions of recovery speed and reliability of the recovered data.  These performance issues may demand the development and implementation of recovery strategies at the point of backup.

However, administrators must remember that costs of recovery increases as the RTO and RPO numbers go closer to zero.  Therefore, there is a need to prioritize on data recoveries and must place appropriate data sets in the Cloud server keeping in mind the recovery protocols that may be instituted.  It should be further remembered that disaster recovery does not demand immediate recovery of all data. The organization invariably is keen to recover current data and may be willing to wait for recovery of other kinds of data, including historical data.

It follows that the administrator must have clarity regarding the organizational recovery objectives and existing corporate data demands and conduct a business impact analysis for inaccessibility of mission-critical and non mission-critical data.  They must work with the IT teams to arrive at practical RTO and RPO for mission-critical data and orchestrate complete recovery of all data over extended recovery times and points that may be specified.

Data recovery plans, in context, must include replication and migration methodology planning.  This process will involve planning for how the data will be backed up, replicated and moved across the WAN when necessary.  The migration policy will have to work with tools that will facilitate this process and the tools will have to be planned for and provided.

Finally, the process of recovery will have to be repeatedly tested, and again tested, to ensure that it works just as it ought to.