When it comes to backup verification, IT management has two competing concerns:
- Management of risk (including cost-of-failure and mean time to failure (MTTF) estimates provided to the business owner).
- Cost to business of maintaining backups.
If you know the cost of maintaining backups exceeds the value of the data being maintained, then you can relax retention rules. The business should adjust contractual obligations, real or assumed, as needed. If the loss of data significantly compromises the business or is potentially a business-ending event, then analysts should carefully weigh the cost of retention and retention maintenance against contractual obligations to the owner or customer.
Considerations
Consider the following when assessing these concerns:
- The cost of rebuilding the data set from scratch.
- The cost of silent failure prevention.
- The overall revenue indirectly and directly generated by the data.
- The profit margin of the data.
- The cost to business due to loss of reputation in the event of an unrecoverable failure.
- Any legal obligations regarding data retention and reasonable retention efforts.
- Insurance policies may cover loss of IP (good backup strategies consider the amount of coverage provided by insurance).
In addition, ask yourself these questions:
- If the data become irrecoverably corrupted, can it be rebuilt? Is the data irreplaceable, such as photographs or videos?
- Are you dealing with real-time decision data? How many hours of human decision data are invested in creating the data?
- Assume the worst-case scenario: You have corrupt, irreparable data that can’t be rebuilt. What is cost to business of data loss?
Summary and conclusion
Establishing reasonable practices is key to any successful backup and data management effort. First, you must develop practices, document these practices, and ensure that IT personnel follow them. Then, you must understand MTTF and maintain a tolerance.
To achieve the highest level of reasonable data retention and silent failure prevention, consider retaining two copies of the data at two geographically distinct locations. Perform weekly MD5 hash checks on the data primary set and monthly checks of the secondary. You can automate these tests and run them in addition to the initial consistency checks performed when the data was created. Developing sound backup and data retention procedures facilitates the safety of data, compliance with business needs for point-in-time recovery, and potentially regulatory compliance. Most database technology includes the ability to restore after a disaster. As such, the performance of the database depends on understanding the way in which logging and backups operate and impact performance. Improperly configuring backups can result in unanticipated outages and even loss of data if backups are corrupt.
For assistance with any of these configurations, contact Relativity Support.
