Opinion: Protect, *then* Perfect – An Inverted Approach to DR
Ed Note: The following is an opinion post by Shannon Snowden, Senior Technical Marketing Architect at Zerto.
One thing has been made clear to me while designing and implemention DR solutions like VMware SRM and Zerto over the past few years: disaster recovery planning and implementation is much more difficult to do than is necessary.
It is thought to be so difficult that it remains substantially behind other datacenter solutions that have leveraged virtualization to increase efficiency and reliability while decreasing administrative effort.
The way DR is still being approached:
- Is based on physical hardware and how hard it has historically been to design and deploy successfully
- Is done from a risk acceptance position instead of risk avoidance
- Accepts Recovery Points (RPO) and Recovery Times (RTO) that allow for far too much data loss and downtime than necessary
- Puts slow-moving organizational dependencies in the critical path for implementation
If you look at a survey that Zerto conducted recently, the results certainly reflect what I’m seeing:
Cost and complexity are both the biggest concerns with ‘difficult to manage’ coming in close third. Even the companies that have a DR implementation, only 23% are confident their DR will work in the case of a real emergency.
Data center outages are frequent and companies find themselves ill prepared:
- At least 76% of respondents have experienced an outage in the past year
- 42% of respondents experienced an outage in the last six months and 86% of those incidents were caused by something other than a natural disaster
- 26-50% of the applications in data centers are considered business-critical and yet 7% of companies have no DR plan at all
- 77% are not fully confident their DR plan will work
I’m not surprised by these results considering the recommendations coming from the conventional wisdom in the industry as well as what I’ve seen myself.
It’s time to rethink DR. It can be much simpler and robust with the technology that is available now.
Let’s simplify things. Focus on getting the machines protected first, then do all the necessary planning after the machines are in a protected state. This in no way negates the need for complete planning for DR, it just changes the order of the process, because technology allows for this inverted DR implementation approach.
There are really only four steps to getting disaster recovery right as shown in this new infographic designed with the help of Scott & David here at VirtualizationSoftware.com:
Protect Your Virtual Infrastructure from Disaster in Four Easy Steps
This is a much better position to be in as a business if an emergency occurs during the DR planning phase that comes at the end of the Protect, then Perfect style of implementation.
Properly implemented, the biggest concerns shown in the surveys are mitigated and allow the organization to focus on the non-technical areas of the Business Continuity and DR plan.