On average, businesses waste about 35% of their cloud spend due to inefficiently using their cloud resources. This amounts to more than $10 billion in wasted cloud spend across just the top three public cloud providers.
Although the unmatched compute power, data storage options and efficient content delivery systems of the leading public cloud providers can support incredible business growth, this can cause some hubris. It’s easy to lose control of costs when your cloud provider appears to be keeping things running smoothly.
To stop this from happening, it’s essential to adopt a new approach to how we manage – and optimize – cloud spend.
It’s not an easy thing to do, as pricing structures can be complicated. However, in this post, we’ll look at how both AWS and Azure structure their pricing, and how you can best determine what’s right for you.
Different types of cloud pricing schemes
Broadly, the pricing model for cloud services can range from a pure subscription-based model, where services are charged based on a cloud catalog and users are billed per month, per mailbox, or app license ordered. In this instance, subscribers are billed for all the resources to which they are subscribed, irrespective of whether they are used or not.
The other option is pay-as-you-go. This is where subscribers begin with a billing amount set at 0, which then grows with the services and resources they use..
Amazon uses the Pay-As-You-Go model, charging a predetermined price for every hour of virtual machine resources used. Such a model is also used by other leading cloud service providers including Microsoft Azure and Google’s Google Cloud Platform.
Another variant of cloud pricing is an enterprise billing service. This is based on the number of active users assigned to a particular cloud subscription. Microsoft Azure is a leading cloud provider that offers cloud subscription for its customers. Most cloud providers offer varying combinations of the above three models with attractive discount options built-in. These include:
What free tier services do AWS and Azure offer?
Both AWS and Azure offer a ‘free tier’ service for new and initial subscribers. This is for potential long-time subscribers to test out the service before committing for the long run.
For example, you can utilize EC2 and EBS on the free tier to host a website for a whole year. EBS pricing will be zero unless your usage exceeds the limit of 30GB of storage. The free tier for the EC2 includes 730 hours of a t2.micro instance.
Azure offers similar deals for new users. Azure’s services like App Service, Virtual Machines, Azure SQL Database, Blob Storage and Azure Kubernetes Service (AKS) are free for the initial period of 12 months. Additionally, Azure provides the ‘Functions’ compute service (for serverless) at 1 million requests free every month throughout the subscription. This is useful if you want to give serverless a try.
AWS and Azure’s pay-as-you-go, on-demand pricing models
Under the pay-as-you-go model, AWS and Azure offer subscribers the option to simply settle their bills at the end of every month without any upfront investment.
This is a good option if you want to avoid a long-term and binding contract. Most resources are available on demand and charged on a per hour basis, and costs are calculated based on the number of hours the resource was used. For data storage and data transfer, the rates are generally calculated per Gigabyte.
Subscribers are notified 30 days in advance for any changes in the Pay As You Go rates as well as when new services are added periodically to the platform.
Reserve-and-pay-less pricing model
In addition to the on-demand pricing model, Amazon AWS has an alternate scheme called Reserved Instance (RI) that allows the subscriber to reserve capacity for specific products. RI offers discounted hourly rates and capacity reservation for its EC2 and RDS services. A subscriber can reserve a resource and can save up to 75% of total billing costs in the long run. These discounted rates are automatically added to the subscriber’s AWS bills. Subscribers have the option to reserve instances either for a 1-year or a 3-year term.
Microsoft Azure offers to help subscribers save up to 72% of their billing costs compared to its pay-as-you-go model when subscribers sign up for one to three-year terms for Windows and Linux virtual machines (VMs). Microsoft also allows for added flexibility in the sense that if your business needs change, you can cancel your Azure RI subscription at any time and return the remaining unused RI to Microsoft for an early termination fee.
Use-more-and-pay-less pricing model
In addition to the above payment options, AWS offers subscribers one additional payment option. When it comes to data transfer and data storage services, AWS gives discounts based on the subscriber’s usage.
These volume-based discounts help subscribers realize critical savings as their usage increases. Subscribers can benefit from the economies of scale, allowing their businesses to grow while costs are kept relatively under control. AWS also gives subscribers the option to sign up for services that help their growing business. As an example, AWS’ storage services offer subscribers with opportunities to lower pricing based on how frequently data is accessed and performance needed in the retrieval process.
For EC2, you can get a discount of up to 10% if you reserve more. The image below demonstrates the pricing of the AWS S3 bucket based on usage.
Comparing Cloud Pricing on Azure and AWS
As the major cloud service providers – Amazon Web Services, Azure, Google Cloud Platform and IBM – continually decrease prices of cloud instances, provide new and innovative discount options, include additional instances, and drop billing increments. In some cases, especially, Microsoft Azure, per second billing has also been introduced.
However, as costs decrease, the complexity increases. It is paramount for subscribers to understand and efficiently navigate this complexity. We take a crack at it here.
Reserved Instance Pricing
Given the availability of Reserved Instances by Azure, AWS and GCP have also introduced publicly available discounts, some reaching up to 75%. This is in exchange for signing up to use the services of the particular cloud service provider for a one year to 3 year period. We’ve briefly covered this in the section above.
Before signing up, however, subscribers need to understand the amount of usage they are committing to and how much of usage to leave as an ‘on-demand’ option. To do this, subscribers need to consider many different factors –
- Historical usage – by region, instance type, etc
- Steady-state vs. part-time usage
- An estimate of usage growth or decline
- Probability of switching cloud service providers
- Choosing alternative computing models like serverless, containers, etc.
On-Demand Instance Pricing
On-Demand Instances work best for applications that have short-term, irregular workloads but critical enough as to not be interrupted. For instance, if you’re running cron jobs on a periodic basis that lasts for a few hours, you can move them to on-demand instances. Each On-Demand Instance is billed per instance hour from time it is launched until it is terminated. These are most useful during the testing or development phase of applications.
On-demand instances are available in many varying levels of computing power, designed for different tasks executed within the cloud environment. These on-demand instances have no binding contractual commitments and can be used as and when required.
Generally, on-demand instances are among the most expensive purchasing options for instances. Each on-demand instance is billed at a per instance hour from the time it is launched until it is stopped or terminated. If partial instance hours are used, these are rounded up to the full hour during billing.
The chart below shows the on-demand price per hour for AWS and Azure cloud services and the hourly price for each GB of RAM.
|VM Type||AWS OD Hourly||Azure OD Hourly||AWS OD / GB RAM||Azure OD / GB RAM|
|Standard 2 vCPU w Local SSD||$0.133||$0.100||$0.018||$0.013|
|Standard 2 vCPU no local disk||$0.100||$0.100||$0.013||$0.013|
|Highmem 2 vCPU w Local SSD||$0.166||$0.133||$0.011||$0.008|
|Highmem 2 vCPU no local disk||$0.133||$0.133||$0.009||$0.008|
|Highcpu 2 vCPU w Local SSD||$0.105||$0.085||$0.028||$0.021|
|Highcpu 2 vCPU no local disk||$0.085||$0.085||$0.021||$0.021|
The on-demand price of Azure instances is cheaper compared to AWS for certain VM types. The price difference is evident for instances with local SSD.
Discounted Cloud Instance Pricing
When it comes to discounted cloud pricing, it is important to remember that this comes with a lock-in period of 1 – 3 years. Therefore, it would work best for organizations that are more stable and have a good idea of what their historical cloud usage is and can fairly accurately predict what cloud services they would require over the next 12 month period.
In the table below, we have looked at annual costs of both AWS and Azure.
|VM Type||AWS 1 Y RI Annual||Azure 1 Y RI Annual||AWS 1 Y RI Annual / GB RAM||Azure 1 Y RI Annual / GB RAM|
|Standard 2 vCPU w Local SSD||$867||$508||$116||$64|
|Standard 2 vCPU no local disk||$622||$508||$78||$64|
|Highmem 2 vCPU w Local SSD||$946||$683||$63||$43|
|Highmem 2 vCPU no local disk||$850||$683||$56||$43|
|Highcpu 2 vCPU w Local SSD||$666||$543||$178||$136|
|Highcpu 2 vCPU no local disk||$543||$543||$136||$136|
Azure’s rates are clearly better than Amazon’s pricing and by a good margin. Azure offers better-discounted rates for Standard, Highmem and High CPU compute instances.
Optimizing Cloud Pricing
Subscribers need to move beyond short-term, one time fixes and make use of automation to continuously monitor their spend, raise alerts for over or underuse of service and also take an automated action based on a predetermined condition.
Here are some of the ways you can optimize your cloud spending:
Cloud Pricing Calculators
Cloud Pricing tools enable you to list the different parameters for your AWS or Azure subscriptions. You can use these tools to calculate an approximate monthly cost that would likely be incurred.
You can try the official cloud pricing calculators from AWS and Azure or a third-party pricing calculator. Calculators help you to optimize your pricing based on your requirements. For example, if you have a long-term requirement for running instances, and if you’re currently running them using on-demand pricing schemes, cloud calculators can offer better insights into reserved-instance schemes and other ways that you can improve your cloud expenditure.
For instance, this Azure calculator by NetApp offers more price optimization option. This includes options to tier less frequently used data to storage objects like Azure Blob and customize snapshot creation and storage efficiency. Zerto is another popular calculator for Azure and AWS with a simpler interface. However, note that the estimated cost is based on current pricing and is subject can be liable to change.
Price List API
Historically, for potential users to narrow down on the final usage cost involved a considerable amount of manual rate checks. They involve collecting price points, and checking and cross-referencing them manually.
In the case of AWS, the Price List API offers programmatic access, which is especially beneficial to designers who can now query the AWS price list instead of searching manually through the web. To make matters more natural, the queries can be constructed into simple code in any language. Azure offers a similar billing API to gain insights into your Azure usage programmatically.
Understanding and optimizing cloud pricing is somewhat challenging with AWS and Azure. This is partially because they offer hundreds of features with different pricing options and new features are added to the pipeline every week.
To solve some of these complexities, we’ve covered some of the popular ways to tackle pricing in AWS and Azure. Here’s a list of things that we’ve covered:
- How the cloud pricing works and the different pricing schemes in AWS and Azure
- Comparison of different instance pricing options in AWS and Azure which includes reserved instance, on-demand instances, and discounted instances.
- Third-party tools like calculators for optimizing price.
- Price list API for AWS and Azure.
If you have any thoughts to share, feel free to post it in the comments.
About the author Gilad David Maayan
Gilad David Maayan is a technology writer who has worked with over 150 technology companies including SAP, Oracle, Zend, CheckPoint and Ixia. Gilad is a 3-time winner of international technical communication awards, including the STC Trans-European Merit Award and the STC Silicon Valley Award of Excellence.
Over the past 7 years, Gilad has headed Agile SEO, which performs strategic search marketing for leading technology brands. Together with his team, Gilad has done market research, developer relations, and content strategy in 39 technology markets, lending him a broad perspective on trends, approaches, and ecosystems across the tech industry.