Backups are necessary, but no one wants to spend excessive time making or managing backups. The moment some critical data goes missing is when you really care about your backups. However, the real value is in the restore. Just like backups, though, all restores are not created equal. A fast and easy restore to production performance is vastly preferable to one that requires a lot of management and takes long enough to cause significant downtime.
This downtime caused by a production system outage and restoration has a definite cost. One study suggests that an enterprise application outage costs up to $9,000 per minute. Being able to restore to production performance in seconds will shorten the outage, reducing the cost of lost productivity or even lost business. SimpliVity hyperconverged infrastructure offers built-in data protection with restores that take seconds to complete. In fact, SimpliVity guarantees that customers can restore a 1 TB local virtual machine (VM) in 60 seconds or less as part of its standard warranty. The restored VMs are at full production performance immediately.
Other backup solutions copy VMs to a separate storage device. The backup storage is often made up of large hard disks that are optimized for large transfers. This keeps costs down and ensures that backups complete quickly. However, these devices don’t provide adequate production performance given they cannot support the I/O profile of typical production VMs.
Organizations have two restore choices: Copy the whole VM from the backup storage to the production storage then power on the VM, or use the slower backup storage and power on the VM immediately. You then can expect poor performance until you migrate the VM to the production storage. Which compromise would you choose? Wait for the restore, or get the VM up fast but with poor performance? Why should you have to choose? SimpliVity lets you have near-instant restore with full production performance immediately.
SimpliVity stores backups of VMs in the same deduplicated store as the VMs. Backups are replicated to other nodes, and optionally to remote data centers, for resiliency. Since the local backup is on the same storage as the live VM, a restored VM immediately has full storage performance. Even the solid-state drive cache is pre-warmed as it contains the unique blocks that were hot. These unique blocks will almost all be exactly the same between the failed VM and its restored backup. The SimpliVity storage part of the VM restore is trivial. All of SimpliVity’s backup, clone and restore actions update a single pointer to the deduplicated data. There is no bulk copy of a whole VM for any of these operations.
With VMware vSphere, the typical restore takes a few seconds, mostly for vCenter to be updated. Then the restored VM is ready to power on. Usually, the slowest part is making the decision about whether to overwrite the broken VM or create a new VM registration in vCenter. The fast restore doesn’t just have a benefit to the business. Because the task is completed faster, IT staff can move on to other tasks sooner. Help desk tickets are closed faster and satisfaction with IT increases.
The faster you are able to restore an application to production performance, the lower the business cost of an outage. SimpliVity allows VMs to be restored to production in seconds, minimizing the business impact of a VM failure.