Costs in the Cloud: How Much Does AWS Cost?

One of the key reasons or benefits for using the AWS cloud is its pay-as-you-go pricing model.  
As Amazon writes on their official website, “AWS pricing is similar to how you pay for utilities like water or electricity. You only pay for the services you consume, and once you stop using them, there are no additional costs or termination fees.” To break this down a bit, I’ll also mention that AWS offers a pay-as-you-go payment option for over 70 of their services. You’ll pay for them individually, as long and as much as you use them.

This model has many advantages. As a manager, you can start developing new projects with no upfront expenses for infrastructure beneath your project or application, and you can scale up or down using only the resources your business needs at any given moment. This way, you have no upfront expenses whatsoever. Paying as you go allows you to be more flexible because you can easily adapt to the ever-changing needs of your business. And, you can use these services without having to sign any long-term contracts or pay licensing fees.

In this post, we’ll show you how to get the most out of this payment model and cut your costs.
To understand what drives public cloud cost and to learn about best practices of cloud cost optimization, take a look at Cost Optimization Strategies for the Cloud Course.

Cut your AWS costs by using the RIGHT pricing model

Image: https://www.slideshare.net/AmazonWebServices/get-the-most-out-of-amazon-ec2-a-deep-dive-on-reserved-ondemand-and-spot-purchasing-models-64032266

One of the best ways to cut your AWS costs is to choose the right pricing models for paying for your Amazon EC2 instances.
Amazon offers four payment methods:

• On-Demand
• Reserved instances
• Spot Instances
• And Dedicated Hosts

It’s important to note that AWS frequently makes new instance types available, so it’s important that we always use the latest generation instance types, as their prices are much better optimized compared to the same instance type of the older generation.
Now, let’s take a closer look at each payment method so you can choose the one that best suits your use case.

The on-demand pricing model is the true embodiment of Amazon’s pay-as-you-go payment philosophy. You have no upfront payments, which means you’re not committing yourself over a long-term period. Amazon will charge you for the compute capacity by the hour, and you can increase or decrease usage depending on your application.

Now, if your business demands flexibility or your application has specific workloads with sudden high or low workload bursts, then this pricing model is ideal for you. You’ll not only have a low-cost solution, but you’ll also have the opportunity to test and assess your needs before going all-in. In most cases, this is best suited for businesses testing their solutions on Amazon EC2 instances for the first time.

Reserved Instances are a great way to save money because you will pay 30% less than what you’d pay for an On-Demand instance. As reserved instances are assigned to a specific Availability Zone, they allow capacity reservation, so you won’t worry whether you have enough capacity to start new instances if needed.

You may wonder why this pricing model is so much cheaper than On-Demand pricing. As the name indicates, these are instances that you reserve up front, which means that you’ll be committing to a one- to three-year period in order to significantly reduce your computing costs.

Three years is a really long period to commit in IT nowadays. A lot of things can and will change—resource demand, pricing—so it’s difficult to design an entire infrastructure for such a long period. Therefore, I generally don’t recommend signing a three-year contract.

While the biggest benefit of this pricing model is surely its lower price, we shouldn’t overlook the fact that it also provides you with a sense of safety knowing that your business has the computing resources available whenever you may need them. Of course, you have to know how your application works, because this is an ideal pricing model for applications that have a steady and predictable computing resource usage.

Here, Amazon went a step further and made this a Standard Reserved instance and added a new reserved instance type: Scheduled Reserved instances.

Scheduled Reserved instances are designed for businesses that run their most important tasks on a periodic basis. Amazon described some of the uses cases for these instances:

• A bank or mutual fund that performs Value at Risk calculations every weekday afternoon.
• A phone company that does a multi-day bill calculation run at the start of each month.
• A trucking company that optimizes routes and shipments on Monday, Wednesday, and Friday mornings.
• An animation studio that performs a detailed, compute-intensive 3D rendering every night.

With standard reserved instances, you can reserve EC2 compute capacity for a period of one to three years, and as we mentioned before, it’s always at your disposal.

What’s different with scheduled reserved instances is the fact that you can reserve instances for a one-year period but only for predefined recurring intervals with a daily, weekly, or monthly schedule. The best thing is, it will usually cost 5-10% less than what you’d pay for an equivalent On-Demand pricing model.

In the table below, we’ll compare on-demand and reserved instance pricing on an m4.large instance type example running on Linux operating system.

NOTE: Instance costs per hour can vary depending on the AWS region.
By looking at the table, we can see that the most optimal pricing model is up front reserved instance pricing. It allows us to save more, while not having to pay for all of the expenses up front.

Spot Instances are very interesting because they entail bidding for spare EC2 computing capacity, and more importantly, you can get computing capacity up to 90% cheaper than the regular On-Demand price.

Spot instances are only for specific use cases such as for urgent computing needs, or for applications that are only feasible to run using very cheap computing capacity. This is a really great way to save on costs because you get computing resources for just a fraction of the original price.

Finally, the Dedicated Host pricing model is basically a physical EC2 server where you can use your existing server-bound software licenses. You have the option to buy it as on-demand or reserved capacity, with the reserved option being up to 70% cheaper.

Choosing the right pricing model for your business use case or application is of paramount importance, and it will allow you to cut your AWS costs significantly, even up to 90% in some cases.

Optimize to reduce costs

One of the methods for cutting AWS costs is using the Reserved Instances pricing model.

Usually, the biggest expense on AWS is for compute resources, i.e. an EC2 instance. For this reason, it’s critical to choose the most suitable instance for our workflow, as the price difference between instance families is significant. Monitoring your resource usage is important because by doing so, you’ll be able to make sure that your compute resources stay unused.

If you need compute resources for just a short period of time and there is no risk of losing your data, you can use spot instances and save up to 90% compared to on-demand instances. This would be the case if you have an application that performs image resizing and uses an EC2 instance for computing and S3 storage for storing the images.

Another way to cut your AWS costs is to simply review them on a regular basis and terminate any unused resources. Although your production instances need to be able to auto-scale to respond to demand, you can shut down your development or testing instances in off work hours or on weekends and save up to 65%.

It goes without saying that any instances you have for training or demos should be terminated after these projects are completed.

Using AWS calculators

Once you start using several AWS services, you can easily lose track of which resources are you using and how.

While there are many ways to save money by optimizing and tweaking your usage of AWS services and resources, this will be useless if you don’t have a clear plan and an overview of your AWS usage. This brings us to the AWS Total Cost of Ownership (TCO) calculator, which is a great tool for predicting and monitoring your expenses. This is an easy-to-use tool that allows you to estimate your cost savings and provides you with detailed reports.

As we mentioned before, assuming and planning your resource usage is very important. Here, you have a dynamic calculator that shows your estimated expenses and savings depending on the assumptions you enter. All you need to do is to enter your current hosting environment configuration, and the TCO calculator will provide you with a detailed cost comparison and estimated savings if you choose to migrate your applications to AWS cloud.

So how much does AWS cost?

When it comes to the AWS cloud, being knowledgeable about pricing models and how best to optimize expense optimization is essential for avoiding surprises at the end of the month.

AWS is constantly lowering the pricing of its services and resources, and we don’t expect that to change in the future. This means that AWS services and resources will be more affordable than ever.

In addition to the tips outlined in this post, I also recommend checking out some third-party tools that can help you better manage and optimize your AWS costs.

To understand what drives public cloud cost and to learn about best practices of cloud cost optimization, take a look at Cost Optimization Strategies for the Cloud Course.