Disaster Recovery, Horses for Courses and Other Metaphors

Just think how exciting the world of disaster recovery has become. What used to be exclusively tape storage has branched out into all kinds of disk storage, virtual snapshots, deduplication and cloud object storage. That’s great for DR managers, right? Not so fast. One of the central elements of disaster recovery is risk mitigation, which has its counterpart in job security – as in ‘I don’t want to get fired because I lost data using a new technology that failed’. The Chinese have an ancient curse that reflects this situation: ‘May you live in interesting times’. Are new disaster recovery solutions a curse for the busy DR manager or a chance to reduce stress while increasing resilience?

Fans of tape point to its cost effectiveness and reliability, and its propensity to continue to beat other solutions on these points. Meanwhile, disk enthusiasts emphasise ease and speed of recovery of specific files buried in the terabytes of backup. Virtualisation partisans say their instantaneous system snapshots are the answer, allowing entire backed-up servers to come on line immediately. For deduplication devotees, the magic is in backing up data without have to store any (or at least, not the duplicate parts). And cloud object buffs explain that their technology allows for seamless, scalable back-up and recovery.

Yet a survey (2013) from Information Week reveals telling comments from disaster recovery managers themselves, such as ‘We are walking the tightrope, with a monkey on our back, on a windy day, over the fast lane’. There may not be any ‘cookie-cutter’ answers. Instead, it will be necessary to painstakingly grind through particular requirements and restraints to select the right horses for the different courses. Disk mirroring for daily storage, virtualisation snapshots for flexible restore points, deduplication and cloud for accelerating to greater resilience and tape for low-cost storage for less critical needs might be one approach. Nevertheless, one thing is sure: practical planning is better than paralysis by analysis.