Last week in New York I had the opportunity to meet several customers at mid-size data centers that expressed frustration at having to deal with DR products chosen by their predecessors. Most CTOs that I met used a myriad of disjointed tools that they had to put together to satisfy service availability, data protection and BC/DR compliance requirements. All expressed frustration at the lack of integration and automation.
After the trip I met with our engineers to discuss the problem and see what reference architecture we could put together to ease the pain and simplify DR while increasing protection. We thought about DR as a production line of discrete functions: scan, hash, identify changes, de-duplicate, compress, encrypt, store and verify – those are the building blocks for the backup of the production site.
We continued now on the DR site or public cloud: replicate and store changes (using either software or hardware replication), restore protected objects, recover-test applications, measure recovery time, certify RTO/RPO compliance, log activity, keep audit traces, and issue reports and alerts.
The DR production line extends protection from individual files to instantly restoring VMs, to failing over an entire virtual data center. At the end of the production line, we reach Disaster Recovery Assurance: the ability to certify that recovery processes are current and operating within SLAs agreed between IT and the lines of business.