What To Do When Your Cloud Service Fades Away

Guest author Sergey Kandaurov is Director of Product Management at Acronis.

Every once in a while, there is a major disruption that impacts many peoples’ lives and company’s operations. Hurricane Sandy and the earthquake events in Japan come to mind. But these are very high-profile events, which are, in fact, rare.

While companies should be concerned with preparing for large disruptions like these, they are much more likely to be faced with a local event such as a water main bursting and flooding their datacenter, or their Internet connection being severed by someone digging in the street.

Using cloud-based services to store data and run virtual machines on a cloud provider’s servers greatly reduces the potential impact of such local events. But that opens them up to another threat: A cloud vendor shutting down a service or even their entire company, such as when Nirvanix announced their company shutdown back on Oct. 4.

Nirvanix was reasonable in giving customers time to copy their data elsewhere, with resources available at least until Oct. 15 to assist in migration—but not every cloud provider will provide that opportunity. When companies shut down, all bets are off and insurance is pretty much meaningless. So what can a company that is dependent on a provider do to protect themselves against this kind of disruption?

The obvious answer is to be less dependent on any given cloud service. In the case of a cloud provider closing their doors, the only safeguard is to have another copy of your data and other virtual machines that you can quickly access. Beyond that, it’s important that this copy can be restored not only to the original system, but to any other type of system, such as from physical to virtual or from cloud to physical. Or from vCloud to Amazon to Azure.

The key here is using backup software with digital imaging. That allows data and machines to be restored to a different kind of system—on-premise or in the cloud.

Any company utilizing cloud providers should be prepared. Setting up a backup system before disaster strikes is the best way to be protected. In this case, Nirvanix’ customers were lucky to have a few weeks of advance notice, so if they move quickly they should be able to backup and migrate their data to other systems.

Here are some pointers should look for when selecting a backup solution.

• Make sure it is able to restore a complete system image, not just data. Re-installing, reconfiguring and re-patching entire systems isn’t easy and is very time consuming.
• Make sure it allows migrating to different platforms. It should be capable of restoring an image of a cloud-based virtual machine to a local physical machine and vice versa. Any-to-any recovery—even machines from different vendors or running on different hypervisors is what is needed.
• Make sure there is a way to tell if the copy is accurate. Look for systems that can verify and have error correcting codes to ensure accurate recovery.
• Make sure the solution doesn’t require servers be shut down to make the copies. The best solution only copies changed data, and takes a quick picture of it rather than slowly cycling through the data.
• Make sure it supports all systems, platforms and service being used, or that could be used in the future.

It’s easy for companies to become dependent on cloud providers, but in the end, the story of Nirvanix here has showed how important it is for companies to hedge against any possible outcome.