This vendor-written tech primer has been edited to eliminate product promotion, but readers should note it will likely favor the submitter’s approach.
Cloud-based backup and recovery can be appealing for various reasons. Large enterprises may see it as an extension of their hybrid cloud strategy, while smaller shops may simply be looking for a "shrink-wrapped" solution. But it is important to understand that traditional on-premise backup procedures do not always apply in the cloud. While the cloud offers simplicity, to some degree that benefit comes in exchange for new concerns.
Here are the key things to consider when looking at cloud-based backup and recovery:
* Think apps, not files. Traditional backup systems tend to be organized around files and folders, but this model isn't a great fit when you're backing up to a cloud. A better strategy is to think of cloud backup in terms of recovering entire on-premises applications. That means backing up the entire application context – including files, data objects, and even executable binaries – so that whichever portion of the application that may have become corrupted can be reconstructed. Depending on your application's design, the most efficient method may be to backup whole virtual machines.
Organizations with deeper pockets may want to couple this with a full disaster recovery strategy, potentially even implementing a full active/active continuous availability strategy. But a comprehensive backup plan can help reconstruct local applications in a pinch at much lower cost than operating redundant active data centers.
* The WAN is not the LAN. It should be obvious that network bandwidth will be a top concern when backing up to the cloud. One of the advantages of cloud-based backups is they can occur throughout the day and not only during off-peak hours, but customers should consider the bandwidth they’ll need and plan appropriately to avoid disrupting other operations.
Installing a parallel fast WAN connection to the cloud provider exclusively for backup is one option, but a costly one. Organizations lucky enough to have collocated their own systems in their cloud provider's will have it a little easier.
For most customers, however, the most effective solution will be to establish a class of network service just for backup, either on a switch or via software-defined networking, that sets a cap on the amount of bandwidth it can consume. The trick is to attack the problem in terms of network engineering, rather than simply throwing bandwidth at it. This strategy, combined with leveraging WAN optimization, data dedup and hypervisor based protocols, seamlessly orchestrated via enterprise-to-cloud management frameworks, can provide cost-effective solutions.
* Who can get at your data? The issue of data sovereignty is an increasingly important one, particularly for organizations with a global reach. The tug-of-war between privacy requirements on one side and national security concerns on the other makes it imperative that companies understand who can access the data they’ve stored in the cloud, when, and under what circumstances.
Sign up for CIO Asia eNewsletters.