How to compare cloud costs between Amazon, Microsoft and Google

The most expensive way to buy cloud virtual machines is to pay for them on demand

How to compare cloud costs between Amazon, Microsoft and Google
Credit: Thinkstock

Comparing prices among the major IaaS cloud vendors is not as easy as simply checking the cost of one virtual machine versus another. Myriad factors influence price: Size of the virtual machine, type of VM and contract length, to name a few.

So how are you supposed to know which vendor offers the best deal? The key is to understand the different offers from Amazon Web Services, Microsoft Azure and Google Cloud Platform and then determine which is best for your use case.

“Comparing cloud pricing is complicated,” begins a blog post from RightScale that analyzes cloud prices among the major vendors. “It can be difficult to make apples-to-apples comparisons because cloud providers offer different pricing models, unique discounting options, and frequent price cuts.”

But by taking a close look at cloud prices, RightScale provides advice for the value shopper.

aws azure google on demand prices RightScale

RightScale found that for on-demand virtual machines, Google and Azure routinely beat Amazon Web Services on price.

How to save money

One advantage of IaaS is that customers can spin resources up and down as needed. But the most expensive way to buy cloud-based virtual machines is to pay for them on-demand. If customers plan and commit to a long-term contract, they will save money.

In AWS, the primary way to do this is by using Reserved Instances (RIs). “You get the discount in exchange for making a one-year or three-year commitment with the longer commitment giving a higher discount,” RightScale engineers explain. “If you also pay for some or all of that committed usage upfront, the discount gets larger.” RightScale says discounts of RIs compared to on-demand VMs range from 24% to 75%.

How should RIs be used? That depends on your workloads. If you have fairly stable requirements, then RightScale recommends purchasing up to 80% or 90% of your VMs as RIs. For workloads with more variable demands, RightScale recommends only 30% to 50% of your VMs should be RIs, with the rest being on-demand.

Google has a different model for incentivizing customers to maximize their cloud use and it’s called Sustained Usage Discounts (SUD). “The SUD, which happens automatically and requires no upfront commitment, gives you a discount on each monthly bill based on the percentage of time that instances in a certain family were running during the month,” RightScale explains.

Basically, the more you use the VM, the less expensive it becomes. If an instance runs for 25% of a month, there is a 20% discount on future use; if its running 50% of the month then another 20% discount is applied. If it runs for 100% of a month there is a 30% discount compared to on-demand pricing.

Microsoft offers Enterprise Agreements (EA) that are negotiated with individual customers, so discount methods are not publicly available. But RightScale says Microsoft sales reps provide significant discounts for long-term usage commitments.

Cost comparison

It’s difficult to say which cloud vendor has the least expensive virtual machines because there are so many different VM options across the three providers. Through an analysis of the options, however, RightScale was able to make some conclusions:

  • If customers use a solid-state memory drive then Microsoft Azure tends to be the most cost-effective option.
  • If you don’t need SSDs, then Google usually comes out as the best deal.
  • AWS is most often the middle-priced option among the three providers.

Once you factor in discounts for longer-term contracts, such as RIs, SUDs and EAs, it becomes even more murky. Google generally has the lowest on-demand pricing for VMs, RightScale finds, while Azure usually meets or beats AWS on on-demand pricing.

Other factors to consider

There are many variables that could impact your pricing, such as where you’re running resources geographically. Domestic operations are generally less expensive than international ones for most providers, for example. If you’re able to separate workloads, then Google and Microsoft offer per minute pricing of workloads, versus AWS, which only offers hourly pricing. Customers should also expect to pay a premium for using a Windows-based virtual machine compared to an open source OS, even when using Azure.

For some customers, cost may not be the top concern. With small price differences among the providers, factors such as developer affinity to a platform and specific tools from the providers that meet your needs can be a bigger deal.

Related video:

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Must read: Hidden Cause of Slow Internet and how to fix it
Notice to our Readers
We're now using social media to take your comments and feedback. Learn more about this here.