TL;DR: AWS Savings Plans are a financial instrument designed to give you significant discounts in exchange for committing to a certain level of compute usage. AWS Savings Plans are a no-brainer compared to on-demand pricing once you have some confidence in your general usage trends. The best part is they don’t require changes to your underlying infrastructure and can offer 70%+ savings in the best case.
What is an AWS Savings Plan?
AWS Savings Plans are a financial instrument offered by Amazon that’s designed to give customers a significant discount in exchange for a certain level of committed usage. Savings Plans were initially introduced in 2019 and in the best case scenario offer up to 72% savings on certain EC2 instances.
That 72% savings is a really big number, right? But getting there requires taking advantage of the unique configuration options available in AWS Savings Plans. In the majority of cases savings are closer to 30% but the exact amount is contingent upon three basic factors:
- What term you’re willing to commit to: 1 year or 3 years.
- What prepayment amount you’re willing to make: none, some or all.
- What compute options your organization uses: EC2, Lambda, Fargate, etc.
So, let’s take a closer look at this pricing model and suss out how it works and define a clear path to that steep discount.
Factor 1: The Term
The first factor you’ll want to consider when making a Savings Plan purchase is the term of your commitment. Ultimately when you purchase a Savings Plan you’re committing to either 1 year or 3 years of usage to get a discount. The longer the commitment you make, the bigger the savings.
In most cases, the discount level you get from going from 1 year to a 3 year commitment doubles the discount rate. As an example, if you commit to one year of an AWS Fargate Savings Plan, you’ll see a ~20% savings versus on-demand pricing. That same exact configuration with a three-year term commitment will see a 45% savings over On-Demand.
Factor 2: The Pre-Payment Amount
The second factor to consider is how much you prepay (if anything) for your commitment. AWS gives you three different prepayment options: None, Some or All. Similar to committing to a longer term giving better discounts, the amount that you’re willing to pay upfront considerably increases the discount versus on-demand prices.
The most popular Savings Plan option we see customers make is a “no-upfront” prepayment which kicks off savings usually at 20%. Doing a partial or full upfront prepayment only marginally helps on a discount rate, typically about an additional 5% savings when you go from committing nothing to a partial, and an additional few percent when going from partial to full upfront.
In general, customers tend to opt for no upfront Savings Plans and in the event there is additional budget they’re looking to deploy, they’ll do a smaller full upfront payment to yield maximum savings on a smaller budget.
Factor 3: What Compute Do You Have?
It’s important to note that AWS Savings Plans are only available for certain compute workloads and have three different options: General Compute, EC2 Instances, and SageMaker.
- Compute Savings Plans: Apply to usage across Amazon EC2, AWS Lambda, and AWS Fargate.
- EC2 Instance Savings Plans: Apply to only EC2 instance usage.
- SageMaker Savings Plans: Apply to SageMaker usage.
Within each Savings Plan type, Amazon applies different Savings rates dependent upon which region you’re running in and what compute types you’re using. Each individual EC2 instance type has a different discount rate which can make the process of even understanding what amount of savings you can yield a challenge in and of itself.
Give Me The Savings
So now that you know everything about how Savings Plans work, how do you realize the savings? Well the nice thing about Savings Plans is that after they’re purchased, Amazon will automatically begin applying the savings to your account. As you use certain computer workloads and different instance types, Amazon will grant the highest discount possible to you which is what Amazon means by being “Flexible”.
As an example, if you have two instances: a C5.xlarge and a M5.xlarge and Amazon grants a 30% savings rate for C5 and a 25% savings rate for M5, they first apply the 30% savings rate to ensure you yield the most savings possible. This means that as your organization makes changes to your infrastructure, you can rest assured that some level of savings will automatically be applied even if you have to switch instance types.