What Causes Epic Fails In Disaster Recovery?

Epic Fail In Disaster Recovery

Recent research has shown that 93 percent of all businesses that lost their data centre for 10 days went bankrupt within just a year. Armed with a few key best practices, you can eliminate this risk by building an effective disaster recovery (DR) strategy that will protect your corporate data and your job.

What causes epic fails In DR? Let’s take a look at the most common reasons DR strategies fail and best practices to overcome these challenges.

1. Malfunctioning Backup Technology

Backups and replication are the foundation of a modern DR plan. It’s critical to ensure that the technology is working as it should – e.g., backups are automatic and reliable, the integrity of your data is upheld through inline and offline validation, auditing is automated via failover virtualisation and backups are replicated successfully to a secondary location.

In simple terms, make sure your data is actually being backed up and replicated to an offsite location. This process provides an added layer of redundancy so that if your primary location’s data centre goes down, information is still accessible via the secondary site, and business operations can proceed unaffected.

Another important consideration to keep top of mind is to never neglect your systems. Data backup is only effective if you have available systems and infrastructure upon which to access it. Consider techniques such as dissimilar bare metal recovery and virtualised ‘instant recovery’ to ensure your data and systems will be back up and running quickly after a disaster strikes.

2. Neglecting To Plan

If backup technology is the solid rock upon which you build your DR strategy, then planning is the accompanying foundation. Strong DR plans focus on three primary areas:

  • People: Identify key operational personnel and provide them with the ability to work remotely or at a secondary location when a disaster strikes. This means giving them direct access to the recovery systems, data and other resources they need to maintain business operations. It’s also important to set up an alternate form of communication in the event that your organisation’s primary communications infrastructure goes down.
  • Infrastructure: Identify key operational infrastructure and make sure its protected. It is incredibly important to remember that the reason that you want your IT infrastructure to successfully survive a disaster is so that your most valuable assets – your staff – can use your systems and your data after the disaster.
  • Processes: Identify key operational processes – step-by-step guidelines for who does what in the event of a disruption – and make sure employees are aware of and have practiced their role. Having cross-organisational buy-in is an extremely important piece of the puzzle. Remember to consider every process that is critical to the daily operations of your business, not just the IT processes.

3. Failing To Test, Test And Test Again

Not performing DR testing on a consistent basis is a major issue for administrators because of the ongoing evolution of IT infrastructure. There is often a divergence that is difficult to capture that is related to, but not wholly reflected by, the assumption of heterogeneity. The cure is thorough, iterative DR testing that is done on a consistent schedule that allows it to be adopted as yet another standard business practice.

Creating an effective DR plan isn’t just about grinding through apparently endless processes and details; it also requires the creativity to think outside the box and consider a variety of different disasters and resulting DR consequences. In other words, seriously consider the disaster scenarios that you deem impossible in addition to those you think most probable – because, in reality, you just never know.

4. Assuming Homogeneity

In order to maximise ROI, it’s important to maintain flexibility so that IT remains responsive and agile to the changing needs of its users and to the evolving technology and vendor landscape. DR failures often occur because administrators support only a single technology in their planning.

For example, they’ll support storage technologies such as SAN or NAS, but ignore direct attached storage; embrace virtualisation, but ignore physical servers; or support physical appliances, but fail to consider the need to restore between different models and generations of servers. Implementing a DR plan and accompanying technology that embraces heterogeneity and adapts to agile IT environments will be the most effective when you need them most.


Implementing an effective DR strategy doesn’t have to be difficult or consume multitudes of IT resources. Functional backup technology, detailed DR plans, continuous testing and heterogeneity is the core foundation that will prevent failures and propel you to success.

SHARETweet about this on TwitterShare on LinkedInShare on FacebookShare on Google+Pin on PinterestDigg thisShare on RedditShare on TumblrShare on StumbleUponEmail this to someone
Mark Campbell

Dr. Mark Campbell is the chief strategy and technology officer at Unitrends. Prior to joining Unitrends, Mark co-founded mindAmp, which provided high-technology business and software development consulting. He has also worked as the senior vice president and general manager of the systems management business at Legent, as well as vice president and general manager of the enterprise systems business at AT&T and NCR.