Asked and answered: Should in-cloud data be backed up on-prem?

Data disasters happen. Cloud services can go down. Fires can hit data centers. Dogs do, on rare occasions, eat homework.

Having a backup copy somewhere helps to avoid surprise data disappearances, of course, but IT pros, especially cloud newbies, may still want to keep their precious info and assets close.

An attendee of an October IT Brew event, “Up in the Air: The State of Cloud Computing,” had the following question for guest David Vidoni, VP of IT at the workflow-automation company Pegasystems:

If you move something to the cloud, should you have your own on-premise backup location?

We posed the question to Vidoni and later, to other industry pros.

These responses have been edited for length and clarity

Vidoni: I think the more important question is: Do you have a backup somewhere—in the cloud somewhere else, or maybe it’s in a different region or a different availability zone? So, if there is ever a problem, you’re able to go to that backup location and get it.

Erik Duffield, CEO and co-founder, Hakkoda: You absolutely need backup-recovery strategies. They do not have to be on-prem…Most cloud solutions and cloud platforms—particularly on the data side, but also on the application side—have multiple layers of point-and-time recovery available to you.

Vidoni: You could just as easily have data be streamed and backed up in another part of the cloud, whether it’s AWS, or Azure, or GCP. I’ve seen some folks put it into an entirely different cloud. They have complete diversity of clouds, so if there’s any major event they can go back to the other one.

Murad Korejo, engineering manager and architect, Presidio: Let’s talk about some of the reasons why [on-prem] might make sense. Firstly, security and regulatory requirements. With HIPAA, for example, you have to do a lot of demonstration and supplying of evidence that [assets and data] are protected and secured appropriately. And that just may be easier to do, in some circumstances, for an organization, on premises. They’re more familiar with that technology perhaps; they’re more comfortable operating in that realm. Maybe the cloud that they’re operating in has some limitations, and there have been issues with trying to meet that regulatory constraint in that cloud.

Duffield: If you take it to on-prem, my concern is you’re moving a lot of cloud to ground and vice versa. And it’s expensive. It’s slow. Is it out of date? It’s harder to keep it in sync. Not that there aren’t people who will choose that for various security or data policies.

Vidoni: My personal approach would be to keep it out [in the cloud], because it’s just easier and cleaner to do it, and you’re not having to worry about hardware, backups, and all that.

Duffield: You can do it. You’re just connecting your on-prem servers to your cloud infrastructure and keeping that pipeline communicating back and forth. If you’re on the cloud, and you’re in a clone situation, it’s happening in real time; you’re not having to maintain those pipelines to your data center.

Korejo: There’s complexity in running an on-premises backup solution, keeping the lights on, making sure it has enough infrastructure and storage capacity. All those things are taken care of for you in the cloud.

Vidoni: If you’re going out into the cloud, I would say, go out; don’t keep one leg on one side, and one on the other. Just make sure that you do have that diversity of where the data is being stored.