Making An Appropriate Budget For Disaster Recovery Systems

August 30, 2018

Tim Annable
Tim Annable MBA, CISSP, CMPE, MCSE Chief Data Scientist

Every business manager faces the struggle of working with their IT staff or vendors to budget for backup and disaster recovery systems.  Many of the components of a good backup system are never utilized and add no value to daily operations.  However, in times of disaster, these same components can be the difference between business recovery and bankruptcy.   Beyond that, there is a seemingly infinite array of backup options and features.  So, how does a manager weigh these budgeting decisions?

First, one must understand the elements of a proper backup system.  The following are the four critical components that make up a good backup system:

1) All data should exist in multiple copies separated in the space dimension. 

  • Separating data copies in geographically diverse locations is a must.  If the backup copies are located in the same location as the primary copies then fire, flood, theft, or other disasters can destroy both copies in the same incident.

2) All data should exist in multiple copies separated in the time dimension.

  • A business should be able to revert to multiple times (i.e. restore a database to the state it was in at close of business yesterday). 
  • Geographically separate copies are not sufficient. If those copies are rapidly replicated any corruption of data could be quickly replicated to the other sites; resulting in two copies of corrupt data.

3)   All data in the organization must be documented and categorized. 

  • Businesses store data in a multitude of places and it can become especially difficult to ensure that all data is being backed up if this documentation is insufficient.

4)    The backup system should be centralized, monitored, and have alerting capabilities. 

  • Backup systems may go unused for years until disaster strikes and can become neglected.  It must be easy for the person responsible to monitor the system.   Alerting, the self-reporting of issues, is also critical to notify the backup operator of issues.

There are many options for backups that drastically affect the capabilities and costs of these systems.  On one extreme, a business might have three redundant data centers – two of which could be destroyed without impact to operations. On the other extreme, a company could have good backups, but in the event of disaster have to order new hardware and have a two-week-long restoration process in which 90% of the data is restored. Either of these extremes may be acceptable based on business needs. 

To assess need, the manager must weigh in two factors:

  • Acceptable data loss window.
    This is the window of time that represents the information that will be lost because it was added to the system since the last backup.  In many businesses, a one-day data loss window is acceptable.  Backups are performed nightly, and the business is prepared to lose (or recreate) a day’s worth of data in a disaster.  In other businesses there is almost no acceptable level of data loss. 
  • Acceptable recovery time window.
    This is the window of time that elapses between the time the disaster occurs and the time that restoration of system functionality is complete.  Some businesses may find it acceptable to have several days of downtime on their systems while data is being restored, while others can only handle minutes or seconds of downtime.

The features required and their associated costs are driven by the above windows, as well as the quantity of data in the environment.  As the level of acceptable data loss goes down, costs rise.  Similarly, as the acceptable recovery time window shortens, costs rise. 

With the business manager thoughtfully identifying the acceptable data loss window and acceptable recovery time window with their team, they can begin to weigh this against the costs of systems.  From here it is simple risk-management math.  For example, if the manager is choosing between a recovery time window of 1 day or ½ day, then a cost difference will be available for this capability in backup systems.  The manager would calculate the NPV of the cost of ½ day’s data loss.  If the NPV of this loss is less than the cost difference between backup systems, it may not be worth the cost.  Conversely, if the NPV of the loss is greater than the cost of the mitigation, it probably is worth the cost.

In summary, the manager must focus on identifying the acceptable data loss window and acceptable recovery time window to gauge an appropriate budget for backup systems.  As long as the systems are budgeted within these windows and contain the four key elements of a good backup, one can rest easy with their budgeting decision.

Try our IT Risk Assessment to begin understanding the magnitude of your organization’s risk.


Back to News