Flexera logo
Image: AWS vs Azure vs Google Cloud Pricing: Compute Instances

Comparing cloud pricing is complicated. It can be difficult to make apples-to-apples comparisons because cloud providers offer different pricing models, unique discounting options, and frequent price cuts. This article dives down into the details to show you which providers have the lowest-cost options for cloud compute instances and under which circumstances.

Understanding Cloud Discounting Options

On average, compute resources represent 75-80 percent of your cloud spend. Before you can accurately compare compute prices, it’s critical that you understand the discount methods for each of the cloud providers since that will be a major driver in the price you pay.

AWS Pricing

The primary discounting method for compute resources on AWS is Reserved Instances (RIs). RIs are not actual instances, but rather should be thought of as discount coupons that can be applied to instances that meet certain criteria (region availability zone, instance family, and operating system). You get the discount in exchange for making a one-year or three-year commitment with the longer commitment giving a higher discount. If you also pay for some or all of that committed usage upfront, the discount gets larger. Convertible RIs are a new type of 3-Year RI that give you a smaller discount, but give you the flexibility to move the discount between different instance families during the term of the RI.

Discounts range from 24 to 75 percent depending on the RI term, the instance type, and the region. Often 1-Year No Upfront RIs are a good starting point, and we used those in our comparisons among clouds. We recommend caution in purchasing 3-Year RIs, since your usage may change drastically over 3 years. Although Convertible RIs allow you to adapt to new instance types, you can see in the chart below that Convertible RIs offer only a 5-6 percent extra discount as compared to 1-Year Standard RIs while locking you in for 3 years instead of 1.

Recommendation: If you aren’t using RIs, you are paying the highest possible price for each instance. Your cloud cost management strategy should define your level of desired RI coverage. Target 80-90 percent coverage for environments with little variability, while 30-50 percent coverage might be appropriate for environments with significant variability.

Azure Pricing

The primary approach to getting discounts on Azure is your Microsoft Enterprise Agreement (EA). EAs offer discounts from 15-45 percent depending on the level of usage you commit to. We used a 30 percent discount as the midpoint in our comparisons below.

Google Cloud Pricing

Google Cloud provides the simplest approach to saving money on compute resources through the Sustained Usage Discount (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. Once instances have been running for 25 percent of the month, the price goes down to 80 percent of on-demand (a 20 percent discount), and when you hit 50 percent and 75 percent of the month, the discount goes up another 20 percent. As a result, instances running 100 percent of the time during the month will earn a maximum 30 percent discount.

Also keep in mind that Google Cloud will combine different instances of the same type when possible to give you the best discount level.

Compute Pricing

Now that you understand the different discount options, let’s compare both on-demand and discounted prices.

Important note: Keep in mind that the Google SUD will apply “no matter what,” while AWS RIs are dependent on your proactively purchasing RIs. Azure discounts will be dependent on your specific EA agreement.

In comparing compute prices, it’s important to note where the compared instances are similar and where they are not. In our analysis, we chose six scenarios to compare, shown in the first column of the table below. We looked at standard, high memory, and high CPU instance types with 2 vCPUs. For each of these three instance types, we looked at scenarios where a local SSD is required and where a local SSD is not required, resulting in a total of six scenarios. All scenarios are based on the lowest price us-east region for each provider and using one of the standard, free Linux distros that are available at no extra charge (not RHEL).

For each cloud provider, we then mapped the appropriate instance type for each scenario. You can quickly see that everything cannot be compared exactly “apples-to-apples”:

  • Notice that the amount of memory varies for the similar instances across the cloud providers. In most cases, the difference in memory is relatively small (10 percent or less), but for highcpu scenarios, AWS and Azure offer more than double the memory of Google.
  • AWS has instances families with and without the local SSD for standard (m3 and m4), highcpu (c3 and c4), and highmem types (r3 and recently introduced r4).
  • Azure always includes local SSD with all of its instance types, so as a result you are “paying for it” as part of the instance price whether you need it or not.
  • Google Cloud never includes local SSD with the instance type, so you need to pay for it as an add-on cost. The minimum size for an add-on local SSD is 375 GB, which is quite large. You can use Google Persistent Disk as an alternative, but it does not offer the same access times as the local SSD.

Comparing On-Demand Compute Pricing: AWS vs Azure vs Google Cloud

For each of the six scenarios below, you can see the hourly on-demand (OD) price for each cloud and then the hourly price per GB of RAM for each. We calculate both so that you can normalize the prices if the amount of memory is important to you.

In the chart below, red indicates the highest price of the cloud providers within a scenario while green represents the lowest price. If there are ties, then both cloud providers are highlighted green or red.

A summary of the results:

  • Google Cloud is the lowest price for 3 scenarios; highest price for 7.
    • Google Cloud tends to be the lowest price when no SSD is needed and the higest when SSD is required due to the price of adding on a 375 GB local SSD.
    • Google Cloud is higher priced on the “per GB RAM” cost for highcpu due to the fact that it includes less than half the memory of AWS and Azure.
  • Azure is lowest price for 6 scenarios; highest price for 4.
    • With the AWS price changes on December 1, 2016, Azure matches or is lower than AWS for only 7 of the 12 scenarios.
  • AWS is the lowest price for 3 scenarios; highest price for 1.
    • AWS is most often the middle-priced option if the 3.
  • All of these results will depend on the level of Microsoft EA discount for Azure and on what type of RIs, if any, are purchased for AWS.

Comparing Discounted Compute Pricing: AWS vs. Azure vs. Google Cloud

In comparing discounted prices, we looked at annual (instead of hourly) costs to give a better comparison. This is because AWS RIs require a commit for at least a year, and the Microsoft EA for Azure is a three-year contract (albeit with the ability to adjust commitment levels each year). Google Cloud is more flexible, requiring only a month of usage for its SUD.

Below for each of the six scenarios, you can see the discounted annual price for each cloud and then the discounted annual price per GB of RAM for each. We calculate both so that you can normalize the prices if the amount of memory is important to you.

In the chart below, red indicates the highest price of the cloud providers within a scenario while green represents the lowest price. If there are ties, then both cloud providers are highlighted green or red.

A summary of the results:

  • The discounted pricing here looks fairly similar to the on-demand pricing above.
  • In two situations where AWS and Azure had equivalent pricing for on-demand scenarios, AWS came out lower priced with discounts applied.
  • All of these results will depend on the level of Microsoft EA discount for Azure and on what type of RIs, if any, are purchased for AWS.

The Impact of Per-Minute Charges on Cloud Pricing

Azure and Google Cloud have touted the fact that they bill on a sub-hour basis. Here’s how each bills:

  • AWS bills one full hour starting from the first minute of use.
  • Azure bills 1/60th of the hourly cost for each minute used.
  • Google Cloud bills 1/6th of the hourly cost for the first minute and then bills per minute after 10 minutes.

If you have a number of workloads that are commonly running less than an hour, per-minute billing could be a material savings as shown in the chart below.

However, for workloads that are running longer than an hour, the impact of per-minute billing quickly disappears, and the actual hourly price of the instance is a larger factor.

The Impact of Regions on Cloud Pricing

Different cloud regions have different pricing for each cloud provider:

  • AWS and Azure have different prices on a region-by-region basis. The most expensive regions in Asia Pacific can be 50 percent more than the lowest-price regions in the U.S.
  • Google Cloud has the same prices for all U.S. regions and an upcharge of 10 percent for all regions in Europe and Asia. As a result, Google Cloud may gain a further price advantage in Europe and Asia.

The Impact of Windows on Cloud Pricing

Each cloud provider charges a premium for instances running Windows in order to cover the licensing costs for Windows.

  • For on-demand, Azure has a slightly higher premium for Windows than AWS.
  • For discounted, Azure has a much lower premium than AWS.
  • In both on-demand and discounted, Google Cloud has the lowest premium for Windows of the three cloud providers.

The Bottom Line on AWS vs Azure vs Google Cloud Pricing

Your prices on compute are going to depend on your particular use case and your discounting approach with AWS and Azure, since Google Sustained Usage Discounts are automatically applied.

  • If you are not using AWS Reserved Instances and don’t have a Microsoft EA discount for Azure, Google Cloud is going to be cheaper in most scenarios because the Google SUD applies no matter what.
  • If you commonly need the performance requirements of local SSD (vs attached storage like Persistent Disk), you’re going to pay a premium for it on Google Cloud.
  • With the December 1, 2016, price cuts from AWS, Azure no longer matches or beats AWS on all prices for on-demand. Expect to see Azure match or beat those prices soon.
  • AWS is likely not going to be the cheapest in most of these scenarios but it is often in the middle.