A data backup trend that is beginning to take hold is converged data protection. The technology can be thought of as similar to hyper-converged infrastructure. The difference is that, while hyper-converged infrastructure systems serve as platforms for running virtual machines, converged protection products trade the hypervisor for integrated backup software. The result is a multi-node backup appliance with its own scale-out storage.
While there are two clear leaders in the emerging converged data protection market -- Cohesity and Rubrik -- they are no longer the only players in this space. Hewlett Packard Enterprise, for example, offers a product called HPE Recovery Manager Central. Asigra also offers converged data protection appliances that are designed for managed service providers.
Three key considerations
Because each vendor has its own way of doing things, those considering a converged data protection platform should carefully evaluate the following product features.
- The software. Some vendors provide purpose-built backup software that is tightly integrated into the hardware appliance and highly optimized. There are also vendors that provide a converged product based on off-the-shelf backup software. These offerings may be a good option if an organization wants to maintain compatibility with the backup software it already uses, but they are unlikely to deliver the performance of a purpose-built product.
- Scalability. This is one feature that makes converged data protection so attractive. However, there is often an upper and a lower limit to this scalability, so it is important to ask potential vendors about minimum size, as well as the upper scalability limit.
- Backup policy flexibility. A good backup product should allow you to specify backup frequency and backup retention by data type or source. Similarly, it is a good idea to ask a backup vendor specific questions regarding how backups are created and stored. Converged data protection vendors commonly use snapshots as a part of their overall approach to data protection, but there can be significant differences in how snapshots are created.
Other must-have features
Ideally, a converged data protection vendor should use distributed redirect-on-write snapshots, rather than differencing disk snapshots. Differencing disk snapshots tend not to scale very well because read performance begins to degrade as differencing disks accumulate. Scalability is one of the primary advantages to using a converged protection platform, so it is important that the snapshot mechanism be able to scale as the size of the cluster increases.
Snapshots should also be application-aware. Because converged data protection products are classified as backup appliances, the software should use Microsoft Volume Shadow Copy Services to create backups that are application-consistent, not just crash-consistent. Even so, it never hurts to confirm that the vendor is creating application-consistent snapshots, and to verify the applications that are supported.
Products should provide instant recovery capabilities for virtual machines (VMs). Instant recovery works by mounting the VM directly on the backup appliance. That way, the VM can be used while a more traditional restoration completes in the background.
Converged data protection products generally consist of multiple server nodes. While these nodes exist primarily for providing fault tolerance and scalability, it should be possible to harness some of their computing power for instant recovery.
Although not necessarily a must-have feature, some vendors allow backups to be used in the creation of sandboxed test and development environments. This functionality works very similarly to an instant recovery, except no restoration takes place. Instead, the backup data helps create sandboxed VM copies that can be safely used for testing purposes.
Converged protection offers capabilities that are similar to more traditional continuous data protection products, but with the added benefit of easy scalability. Because the converged data protection market is still relatively new, there are significant differences between the vendors. As such, it is important to assess vendor costs and capabilities before deciding on a final product.
Q&A discusses evolving converged data protection market
Adding automation to a converged data protection platform
Hyper-converged infrastructure tackles secondary data storage