peshkova - Fotolia
Recovery-in-place has emerged as an alternative to traditional backup for faster data restoration times. However, the technology has some limitations. Typically, in-place recovery takes approximately 15 minutes. If that is not fast enough, replication will likely be required.
With the replication technique, the same software and process is used, but the secondary image is updated live, typically on a separate storage system. A host is often attached to the storage system with "warm virtual machines" (VMs). This allows for a near-instant (less than 10 minutes) recovery of critical applications, since the VM merely needs to be booted in the case of a failure.
Failback from replication is equally quick. Data is essentially reverse-replicated from secondary storage to primary storage. Once that process is complete, the application is typically shut down, a final quick sync is performed, and the application is restarted and pointed to the restored data on the primary storage system.
The downside to replication is the additional cost of a secondary storage system that can provide performance nearly as good as the original production system. The secondary system needs enough capacity to house the data from mission- or business-critical systems that need a sub-15-minute recovery.
Another downside to the majority of data backup platforms that provide integrated replication is that they do not allow their replication product to provide historical recovery. Replication is essentially for rapid access to the most recent copy of data, not pursuing various Changed Block Tracking jobs. This means the organization will need replication for rapid recovery and backups for 15 minute-plus recovery.
The most important part of a rapid and accurate data recovery has to do with practice. Testing the ability to recover an application is the most important aspect of the data restoration process. What delays most recovery processes is testing. The good news is that some data protection products can leverage the above described recovery-in-place technique to create an on-the-fly "lab" with its own network so applications can be recovered and fully tested without conflicting with production systems.
Data restoration using copy data management
Snapshots are a backup alternative that have been available for more than a decade. Most storage systems now support hundreds of snapshot copies without any significant performance penalty. However, snapshots have a limited ability to assist in the finding and restoring of discrete data. Essentially, most snapshots are only usable to recover the last known good copy of an entire volume, not a subset of data that may be a couple of days old. A potentially larger concern is that snapshots are exposed if the primary storage system that created them fails. If this happens, all snapshot data is lost.
Backup products have begun to address this challenge by integrating with the storage system's snapshot function, a capability sometimes called copy data management (CDM). In addition, there are new products coming to market that specifically focus on this capability.
Faster RPO made possible through
evolution of backup
Though a recent survey by Veeam showed organizations want a faster recovery point objective (RPO) and recovery time objective (RTO) from their IT teams, it wasn't until backup products evolved to meet those windows that it became possible.
The study indicated that most users expect to recover crucial applications in faster than two hours, while most IT staffers said they cannot deliver recovery sooner than four hours.
Indeed, to meet these tighter RPO and RTO windows requires that secondary data be positioned in a way that allows the application to access it faster. This requires the reduction, or outright elimination, of the amount of data moved back to production storage.
CDM indexes snapshots as if they were separate backup jobs. This facilitates search so older data subsets can be found. These backup products can also integrate with the storage system's replication functionality to make a copy of the snapshot to a secondary system. This data restoration capability protects against snapshot data loss if the primary storage system fails.
Finally, some CDM offerings present the secondary snapshots as live volumes. In the event of a failure, a snapshot can be mounted, set to be writable and an application can use the secondary copy until the failure can be repaired. Some CDM products extend this capability so they can use secondary data to drive application testing/development, archiving, e-discovery and analytics processing.
Meeting user expectations of a sub-two-hour backup window is possible. Virtualization makes it easier to meet tight recovery point and recovery time objectives. Data protection tools that integrate snapshot, replication and traditional backup allow users to take frequent and rapidly accessible backups. Once these selections are made, it is important to continually test recoveries.
Convergence of backup and disaster recovery has changed the face of data restoration
Disaster recovery mistakes to avoid when crafting a data restoration plan