Hey everyone, Matthew Sanabria here. I want to talk about this move that we’re seeing from cloud computing to on-premises computing. And if you’re familiar with this already, you might’ve heard the term cloud repatriation. Now I’m not really a fan of that term because I don’t think it accurately describes what’s happening here, but that’s another discussion for another time.
Why are we seeing companies move from cloud computing back to on-premises?
Well, there’s a couple reasons here. First and foremost, cost, right? The cloud environments, they’re on a rental model, a pay per usage model, and that cost can be higher than you’ve anticipated or budgeted because of egress costs or just usage costs.
So if you have a more steady workload, that’s not so peaky and dynamic, on-premises probably makes sense for that from an economical cost perspective. If you were to look at probably every company out there, odds are they have a cost controlling project on their roadmap, and part of this is because they’re seeing the migrations to the cloud wasn’t as cost effective as they thought it was gonna be.
Outside of cost though, you’re seeing things like regulation and jurisdiction and data sovereignty where there are requirements to run certain workloads at sites or locations that you either own or are within certain borders. On-premises is really suited to that, especially if your cloud providers don’t have regions to cover this. So we’re seeing companies bring their workloads back for those data sovereignty concerns.
We’re also seeing things like the performance side of the house, where when you rent computing in the cloud, sure you can pay more for larger instances or more resources, but you don’t truly own all of the hardware there or all the resources.
You don’t have access to all the IOPS, you don’t have access to all the CPU, you don’t have access to the hardware. So we’re seeing companies move those workloads that are performance sensitive or even just latency sensitive to be next to something that is, you know, some dependency that’s in the data center. We’re seeing companies move those back on-prem on their own hardware and that gives them the performance they’re looking for.
So these are kind of those three pillars that you’re seeing companies look at. And these can be further subdivided into further, you know, specific pillars. But these are sort of the three pillars we’re seeing out there: cost, data sovereignty, and performance. So I just wanted to kind of talk about that a little bit and see what you have to say. Has your company moved their workloads from the cloud back to on-premises? What was your experience like? Why did you do it? Let me know in the comments and I’d love to chat about it. Thanks.