The age of application-centric solutions has finally begun with the rise of hyper-converged infrastructure and enterprise cloud solutions. The key ingredients responsible for this milestone are effective compute and storage-level abstraction with highly scalable architectures compared to traditional, legacy three-tier infrastructures. One could expect the backup solution that is responsible to protect their same applications would be as efficient as their production counterparts. But, in reality, this couldn’t be further from the truth as most backup vendors are only relevant to their traditional, legacy three-tier data centers. From our perspective, it starts by asking a simple set of questions to any enterprise backup vendor you may be considering:
Question: Can the backup solution auto-discover all business-critical applications across your datacenter, regardless of location?
Discovering VMs is pretty much a simple operation, as the parent cluster provides the necessary APIs. However, discovering the underlying applications within the VM is a totally different ball game. This is because the application resides on top of the OS stack and every application is unique. Hence, the burden falls on the shoulders of the end user to identify VMs running a specific application and then run a manual discovery process as per the recommendation of most backup vendor’s solution on every VM. Is it too much to ask for the solution to scan and discover the underlying applications across all the VMs in a single stroke?
Question: Can the backup solution assure you the recovery time in case of a disaster?
While RPO has always been a metric of assurance and accuracy, Recovery Time Objective (RTO) has always been a metric of educated guesswork. Experts always try to come up with an approximate number based on the amount of data to be recovered, available network bandwidth and throughput. Just like RPO, RTO is a crucial metric for a business to successfully survive a catastrophic failure. An ideal application should have the ability to constantly monitor the backup’s data transfer and throughput and maintain historical data so that it can help assure the required RTO of a given dataset.
Question: Can the backup solution guarantee application consistent backups with no VM Stuns and no performance impact?
The biggest challenge of performing a backup of an application is not just about guaranteeing its recovery, but also to ensure zero negative impact on the production copy. IT admins do everything they can to avoid performance impact on production applications by scheduling backups during off-production hours thereby compromising RPO. Most backup vendors claim it is an unavoidable problem, which can’t be further from the truth. This can be mitigated if the solution has a solid groundwork in leveraging the underlying platform and enforce best practice mechanisms to protect a given application.
Question: Can the backup solution avoid the pain of deploying and maintaining agents and plugins for all applications?
Most backup vendors provide agents or plugins to be locally installed in each VM to discover and protect these applications. This can be a time-consuming operation and painful to maintain and upgrade. The true measure of a solution’s elegance lies in its level of intuitiveness to solve a given problem. What if there was solution that didn’t need any agents or plugins and could "automagically" protect all the applications within the VMs via remote procedure calls?
Question: Can the backup solution offer a simple, fast, granular recovery of your business-critical applications or database?
Traditional backup solutions can offer granular restores with the help of agents and plugins deployed on individual VMs. Despite all that, the recovery times are still dependent on the backup target and the available network bandwidth. Furthermore, the requirement of agents for granular restores can complicate the entire recovery workflow which is a time consuming and frustrating experience. All this could be mitigated, if the solution could leverage the underlying host platform with an agentless architecture. This could not only aid in fast and granular application recovery, but also provide instant application clones and fast mount capabilities.