Data protection becomes more challenging as enterprises become more distributed. Data is distributed across core and edge data center environments, and on and off-premises applications and infrastructure resources. This data must be readily accessible by users regardless of their location or from what device they are attempting to access it from. Meanwhile, service level agreements (SLAs) for recovery point and time objectives (RPOs and RTOs) and data privacy regulations are only becoming stricter.
The cloud can be used to address the leading problems that are associated with protecting the distributed enterprise with legacy approaches, including high costs, complexities and limited scalability and accessibility. However, if not implemented correctly, the cloud comes with a number of its own potential pitfalls.
Some cloud backup solutions keep process scheduling, management, index creation, storage and deduplication on-premises, and then replicate data off-premises to the cloud. Or, if they do use the cloud for these processes, they do not do so in a scale-out fashion. With both of these approaches, the enterprise is not taking advantage of one of the key benefits of the cloud, which is the ability to pay for compute cycles as they are used. The enterprise is left paying for compute cycles that are sitting idle for the majority of the time.
Another area in which many cloud backup solutions are lacking is in the ability to tier older backup copies to the cloud. They simply use the cloud as a secondary mirror copy, and all data must still be retained on-premises. This approach is expensive, and effectively causes the enterprise to double pay for storage. The enterprise is still paying to own and manage the on-premises infrastructure that is required to store all of this data. This creates not only a budgetary challenge, but also the challenge of finding available data center floorspace in which to host this capacity. Additionally, the enterprise is also paying to host that data in the cloud (and on a tier of cloud storage that is most likely not the least expensive tier).
The ability to recover either in the cloud or on-premises is also important. Each recovery job will have varying amounts of data that needs to be recovered, as well as varying requirements around how quickly that data must be recovered. Recovery to the cloud can potentially add levels of cost and latency that are not acceptable for certain backup jobs.