What you will learn from this tip: With the increasing availability of cost-efficient products, many view remote data replication as the DR strategy of choice to address tighter RTO and offsite media handling issues. This tip outlines the other considerations you should keep in mind that can help you take full advantage of this technology.
We have seen a significant increase in the number of low-cost data replication products available to small and midsized (SMB) IT departments in the past few years. Given the technology's enhanced performance from a backup, and, even more, a restore perspective, many see it as an efficient and cost-effective answer to tighter recovery time objectives (RTO). Decreasing bandwidth costs have also contributed to the technology's gain in popularity.
The ability to quickly access an online (or nearline) copy of the data at an alternate location, the elimination of lost or damaged tape media typically associated with handling and the reduction in manual interventions are all factors that make remote data replication an appealing disaster recovery(DR) strategy.
However, this rapid recovery shifts our attention to other considerations that may not have been as high on the priority list with traditional tape backups. The list below can help you to take full advantage of this technology:
Data categorization: An organization would likely not enable remote replication for all data for cost-saving reasons. This introduces the need to identify business-critical data and assign priorities. In essence, this is the development of service levels or tiers.
Data access: With traditional tape backups, data access issues could be addressed while data was being restored to a replacement system. The ability to almost instantaneously access a remote copy of the data means that applications and users must be redirected to that data just as quickly. This implies some planning around network path redirection.
Data synchronization: You should identify dependencies between various data sets during categorization. Replication may allow a recovery point that no longer matches that of other data backed up to tape, thus introducing data synchronization issues. It may become necessary to promote certain lower criticality systems to a higher priority level based on interdependencies.
Documented procedures: Remote data replication is definitely the DR strategy of choice but does not constitute a DR plan (DRP). No matter how seamless and automated the replication, configuration and procedures must be documented and integrated with the DRP and, ultimately, the business continuity plan. Maintain documentation on configuration and procedures documentation on a regular basis -- or when significant storage changes are made -- so it remains current and relevant.
Monitoring, notification and testing: As with any other backup procedures, you should implement monitoring and replication failure/success notification. Validate the replicated data on a regular basis to avoid unpleasant surprises. Test the system's ability to acquire replicated data sets regularly to ensure that any issues or undocumented changes are captured and remedied.
You should also consider the data protection requirements at the remote location. In the event of a disaster causing the loss of the production copy of the data, the remote replica is promoted to production status until the primary site is restored or repaired, which can take quite some time. This replica is now the only production copy of the data and should therefore benefit from some form of data protection to reduce the risk of data loss. A corporate decision must be made with respect to the service levels for data availability while the organization operates in recovery mode.
Like anything else, a recovery strategy is only as good as its weakest element. Without some forward thinking and comprehensive planning, remote replication is like driving to an unfamiliar destination without directions; you might get there…eventually.
Do you know…
How to troubleshoot your D2D2T system?
How to prepare for next-generation backup?
About the author: Pierre Dorion is a certified business continuity professional for Mainland Information Systems Inc.
