When determining the true cost of your data protection infrastructure, a key factor is that infrastructure must be purchased upfront and in bulk, and requires significant management overhead. These costs can quickly break the bank in a world where more data must be retained for longer periods of time, and where applications have increasingly unique and specific data protection requirements. In this chapter, we will take a closer look at the converged data protection solutions that are marketed as cutting many of these costs.
Converged data protection solutions converge in two ways. The first is that they converge multiple data protection capabilities, including archive, backup and disaster recovery, onto a common software platform. The second way they converge is that they pre-integrate this software onto underlying nodes of compute and storage.
The good news for storage managers is that converged data protection solutions typically provide a scale-out, as opposed to scale-up, architecture. The data protection implementation is no longer confined to the compute power or the storage capacity limitations of a single system. However, these solutions still require a heavy on-premises footprint. This means that on-premises infrastructure still needs to be planned for and purchased. It does not solve the fundamental problem that capacity must be overprovisioned from the start, to accommodate future growth. Overprovisioning occurs even in scale-out architectures because even though they expand with the addition of a node, adding those nodes takes time and rearranging of data center rack space.
Overprovisioning becomes a particular challenge given the unpredictable nature of protection storage. Multiple copies of production data might be spun up to support a range of business use cases such as test and development; capacity needs tend to scale much faster than those of production storage, and the storage manager does not typically have full visibility into what those requirements will be in the future. Converged data protection strategies also do not address the challenge that data protection is frequently an afterthought that is not planned for upfront.
In addition to the cost related to infrastructure itself, the converged data protection infrastructure still needs to be deployed, upgraded and managed. There is typically a perception with scale-out infrastructures that nodes can simply be dropped in as needed, but adding a node requires a number of tasks that take considerable time. For example, the IT professional needs to find available data center floorspace (not an easy task in most data centers today), and also needs to network that node into the existing infrastructure. Once the infrastructure itself is stood up, backup jobs typically will need to be balanced. All of these processes expose the enterprise to the potential for human error and require a significant amount of the IT professionals’ time.
A final challenge, relating to on-premises converged data protection infrastructure, is the lack of the ability to leverage cloud compute and storage resources as they make sense according to cost and performance parameters. The customer is locked into the specific architecture. Each application and each enterprise today have divergent data protection requirements, creating the need for more flexibility.
An alternative to traditional on-premises infrastructure and recent converged architectures is software-as-a-service (SaaS) data protection solutions. Data protection-as-a-service solutions offer similar promise as converged solutions but still have their shortcomings. Until then, access our on demand webinar with Actifio, “How to Create an Infrastructure-less Backup Strategy,” to learn more about how to overcome challenges related to modernizing data protection.