It is nearly impossible to create the perfect disaster recovery plan, one that will work in every situation and...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
meet all your recovery objectives. Instead, the disaster recovery planning process is ongoing and needs continual refinement. One area that we often see as a weakness is the relative distance between your primary data center and the DR location.
The proximity problem
It may seem like a DR site that's close by your primary data center is ideal. The cost of a high bandwidth, site-to-site connection is less within a metro area. It also makes testing easier, since you can leverage that same high bandwidth or simply drive across town to perform tests.
Convenience and cost are key factors, but it is important to remember that outages vary widely in severity. Depending on the nature of the disaster, a nearby site may provide adequate protection. For example, a crashed server or even a whole data center outage is far more common than widespread disasters such as a severe weather event or other region-wide natural disaster. However, a single regional event can do more to impact the long term viability of the organization, to the point of the organization closing its doors forever.
The regional outage reality
There are multiple issues that can impact your DR planning process, if your DR site is located within the same region as your primary data center. Number one, of course is the fact that the disaster itself also impacts your DR site. Even if your DR site is hardened, it will be operating under stressful conditions; for example, using diesel generators for power. Worse, it could be knocked out completely.
Second, the ability of your employees to access the DR site will be severely hampered, since the streets and public transportation routes are likely to be impacted.
Third, in a regional disaster employees tend to focus on their families before they worry about the organization, and rightfully so. This includes employees at the DR site, for example, if you are renting space in a colocation facility. Those employees may not be there to let your employees in if they do manage get to the location.
Also, many colocation facilities "oversubscribe" their capabilities. In a regional disaster this can cause significant issues, because many organizations will need access to the facility at the same time.
Developing a geographic recovery mindset
Because of these concerns it is critical that your disaster recovery location be a fair geographic distance from your primary data center. How far is far enough? A best practice is to make sure your DR site is at least two regions away. FEMA, the Federal Emergency Management Association, provides a map of the US that is broken into regions.
In larger countries like the US and Canada it is fairly easy to identify a DR location that is far enough away. In other countries, especially European countries, establishing distance can be difficult because of the size of the country and restrictive data sovereignty regulations. Until these regulations change, the best design for European organizations may be to establish two DR locations at opposite ends of the country in which they operate.
The tools for meeting a geographic recovery objective
The good news is that the tools to meet a geographic recovery objective (GRO) are significantly better than they used to be. Most storage systems today have native replication and some level of application access. If you have a second facility that is far enough away you could replicate to a second, similar storage system there.
There are also over a dozen third-party replication software utilities available designed to protect specific applications or operating environments like VMware. These applications often allow for any-to-any replication so that the storage devices at each site can be different.
Many of these applications take the process a step further and can replicate to a public cloud provider like Amazon, which can then replicate to multiple sites. These cloud providers will often allow you to start your applications in their cloud in the event of a disaster.
Distance is a critical component of any disaster recovery plan. Despite the advantage of a local disaster recovery location, the risks associated with having both your primary data center and your secondary data center struck by the same disaster are too great. In short, plan for distance in your disaster recovery planning process.
About the author:
George Crump is president of Storage Switzerland, an IT analyst firm focused on storage and virtualization.
10 things your DR planning process must avoid
One thing your disaster recovery planning process must include
Building a foolproof disaster recovery planning process