The recent move by AWS from a per-hour to a per-second billing model, for its Linux-based EC2 instances, was only to be expected, reflecting the increasing demand for scalability in computing solutions. But at the end of the day will it really save you money? Could this increased flexibility in pricing create a too-relaxed attitude, amongst those responsible for procuring this expensive resource?
A long awaited change
AWS’s per-hour billing system used since the EC2 service was launched in 2006, which meant paying for an hour even if you only used an instance for a few minutes, seems almost comic in hindsight, especially for a resource with a reputation for being expensive. The majority of the cloud infrastructure gang, including Google and Microsoft, had since moved to a per-minute model. So it really was high time AWS got a little more precise about charging for some of it’s most popular resources (the new model applies to Linux On-Demand, Reserved, and Spot EC2 instances, more detail on that from TechCrunch). It’s a continuum towards the real flexibility offered by a pay-as-you go model.
Cloud trend: flexibility, scalability
It’s the same trend, it’s the same need that’s driving cloud adoption. Considering the sub-second billing of increasing popular “serverless” platforms like AWS Lambda, charging for anything by the hour, has seemed pretty outdated, and un-user friendly for a while now. It only took the AWS announcement, for Google to immediately follow suit, actually implementing the change before AWS. More on that here. And while we are waiting for Microsoft's announcement, Azure does bill per-second for container instances.
So good news right?
We are getting closer to paying for what we actually consume. Used carefully it will reduce costs. Think of unpredictable workloads that need to be constantly scaled or large one-off tasks. No need shell out for an hour, when you’re only using a few minutes.
And the downside?
Like anything that seems better value or less expensive, we have the tendency to use more of it, more liberally. The thing is, the very important small print is that there is a one minute minimum charge. Remember those prepaid cell-phones in the 90’s, same thing. If you only use it for a few seconds, you’re still charged for a minute. Better than better than an hour, yes. But the “freedom” this offers, can mean developers can spin up loads of large instances under the impression that costs are minimal under a more “user-friendly” pricing model. A scenario that could actually end up costing you more than before.
Lifting the fog surrounding cloud resource management.
IaaS remains an expensive resource and companies still need —in fact now more than ever—to implement some kind of a procurement system to manage cloud costs. Making sure management can see what resources are being used by whom and for what projects. We recommend the cloud resource management platform, OneKloud to bring accountability to your cloud services and lift the fog that too often surrounds their management. It’s the only offer of it’s kind currently on the market and is saving companies between a third and half of their cloud infrastructure budgets.
The impact on you?
So how will this AWS pricing change will impact your business -will you see your AWS bill going up or down? And what needs to happen to ensure this move towards flexibility won’t translate as a free-for-all to your developing teams.
Check out the official AWS announcement for more details, and leave your comments in the box below...