At first glance, understanding cloud computing pricing models can feel like learning to make a grilled cheese sandwich: It's so easy that anyone can do it, right?
Well, not exactly. Although cloud pricing often seems dead simple, the reality is more complex. Much of the confusion stems from the fact that cloud buyers subscribe to various myths that obscure just how complicated it can be to figure out what cloud computing will actually cost you, let alone optimize your cloud spending.
To provide some clarity, let's take a look at common cloud pricing myths, and discuss why cloud pricing is often more complicated than it seems at first glance.
Myth 1: In the Cloud, You Pay Only for What You Use
Most lists of reasons to use the cloud include a line about how in the cloud, you only have to pay for the infrastructure that your workloads actually consume. The idea is that the cloud facilitates much more efficient spending because you don't waste money on servers that you don't need, as you would if you bought physical machines and then ended up not using all of them all of the time.
The reality is more nuanced. It's certainly true that, in the cloud, you can spin up server instances when you need them and shut them down when you're done using them. You pay only for the time the servers are actually running.
However, this doesn't mean that you can't waste money on cloud servers that you don't need. You could keep servers running unnecessarily because you forget to spin them down. You could also choose server instances that are more expensive than you actually need for your workloads. Both mistakes will lead to wasted money. You'd still be "using" the servers in the sense that they would be operating, but you wouldn't be making full use of them.
The bottom line: The cloud isn't a safeguard against the risk of wasting money on infrastructure you don't need. You can overspend in the cloud just as easily as you can on-premises.
Myth 2: Cloud Pricing Is Pay-As-You-Go
You've likely been told that the cloud is wonderful because you pay as you go. There are no upfront costs or major capital expenses to worry about — at least in theory.
In reality, there may well be upfront expenses. If you choose a "reserved" cloud VM instance, for example — which you may because reserved instances cost less — you may end up paying upfront for the instance.
On top of this, there could be major expenses associated with migrating workloads to the cloud in the first place. Refactoring applications, training employees, setting up new tools, and so on could cost hundreds of thousands of dollars or more.
So, while it's true that the cloud sometimes lets you pay as you go in a simple fashion, that's hardly always the case.
Myth 3: Support Services Are Built Into Cloud Pricing
You might assume that, if you're paying a cloud provider for IaaS or SaaS services, support is built in.
Actually, it may not be, at least not to the degree you need. Although the major cloud vendors do offer "basic" support plans for free in conjunction with most of those services, those free plans typically amount to little more than accessing publicly available documentation and automated tools to help manage your environments. If you want actual, hands-on technical support, you'll usually need to pay extra for it.
Myth 4: Cloud Pricing Is Consistent
Perhaps the biggest flawed assumption you could make about cloud pricing is believing that each cloud vendor's prices are consistent.
They also don't go out of their way to make it easy to compare prices between regions, which means it can be very difficult to know you're getting the best price for a given service. Hosting your workload in, say, Western Europe could turn out to be significantly cheaper than in Central Europe, but you won't know it unless you do the comparison manually.
Myth 5: Cloud Prices Are Transparent
All of the major cloud vendors publish complete pricing lists for all of their services. They even offer cost calculators to help you predict how much a given workload will cost you. It's hard to imagine a more transparent pricing model than this, right?
Well, arguably, it's not that transparent. Although it's true that cloud vendors don't hide their prices, they do use exceedingly complex pricing schemes that can make it very challenging to optimize costs. In addition to charging base fees for their various cloud services, cloud providers charge for things like data egress (but not ingress!) and cluster management. Your pricing could also vary dramatically depending on whether you choose auto-managed infrastructure (like "Fargate" mode on AWS) or self-managed infrastructure. And then, again, the prices vary widely between regions.
For these reasons, the way cloud providers publish pricing information seems akin to what is known in the legal world as a "document dump": They deluge customers with so much pricing information and detail that it's hard to sort through it. Cloud pricing may technically be transparent, but that doesn't make it very accessible or understandable.
It's easy to underestimate the challenge of controlling cloud spending. But that's a huge mistake. Predicting cloud costs and optimizing expenses are just as difficult as deploying and managing cloud workloads — even if the cloud vendors don't always admit how complicated their cloud pricing is.
About the authorChristopher Tozzi is a technology analyst with subject matter expertise in cloud computing, application development, open source software, virtualization, containers and more. He also lectures at a major university in the Albany, New York, area. His book, “For Fun and Profit: A History of the Free and Open Source Software Revolution,” was published by MIT Press.