Tips and Tricks: Do You Have a Disaster Recovery Plan?

November 2008

Situation
You need a Business Continuation Plan for an outage, be it a natural disaster like the recent hurricane that caused a wide spread power outage or other catastrophe.  While basic backup and restore plans are adequate for most corruption problems, having a true plan for disasters is essential for continued customer service, and enterprise communications, such as email and personnel management.

Problem

There is no remote plan in place at your company.  Many data centers have a simple backup with no plan as to where to restore the data once the disaster is over.  Many sites have tape backups, shipped offsite which is a good plan, until they must restore to a server that is not there.

Solution
A remote data center site should be part of your plan.  By having a server, or servers available that have the capability of restoring your data from a tape, external disk or internet vaulting service, you will have your customers back up quickly.
The ideal situation for a BCP solution would be to have a second datacenter to replicate data to, and even balance workload.   However, not all firms can afford this type of expense.  
A less costly solution is to replicate data across your campus to servers not on the same power source.  By working with your energy provider (e.g. DP&L, Duke, etc.) they can offer you a differing source.

Tips
Data replication can be done on a frequent basis by your existing SAN storage technology.  The major players in storage all have replication technologies.   Most firms will replicate the data on a nightly basis.  Testing of the recovery scenario should be done at least semi annually, ideally quarterly.  Be sure to have application servers capable of handling 50% of your normal customer traffic at your remote site.  The Exchange servers can be brought up in order of priority.  The database servers should be brought up prior to re-pointing customer applications to the remote site.

Storage providers can provide snap, mirror (asynchronous is recommended to avoid a performance hit), or BCV methods to replicate your data.  The least expensive of these would be to run nightly backups of your database, then replicate the backup files to the remote site.  This will require a restore at the remote site in order to fail over.  It will take time with large databases or exchange storage groups.  If your budget allows, snap technology reduces this time to recovery to minutes.
Therefore, as with all solutions, your budget will drive your time to recovery.  However, the most important item is that you have a backup of your important data at a site other than that where your production systems reside.  Even a low cost backup is a backup and your data is your most important asset.

If you need more information on Disaster Recovery Planning, contact Gary Codeluppi at 937-431-1026 x123 or visit the Ross Group Inc web site at www.rossgroupinc.com.
Comments (0)Add Comment
Write comment
 
 
smaller | bigger
 

busy
search | login