AWS Savings Plans are a type of commitment with slightly more flexibility. Still, every hour when your team uses fewer cloud resources than contracted is pure waste. On the other hand, if you commit too conservatively, you risk under-committing and paying for the necessary resources at the highest AWS pricing tier—On-Demand rates.
Continue reading to learn how AWS Savings Plans work and what else you can do to ensure your cloud infrastructure doesn’t break your wallet.
What are AWS Savings Plans?
AWS Savings Plans are pricing schemes that offer lower prices on regular On-Demand instances in exchange for committing to one or three years of usage. Such a plan requires a one-year commitment to an aggregate per dollar per hour commitment.
You can set your hourly spending restriction, and all compute use is discounted. AWS will charge you the regular On-Demand price when you exceed your allowed threshold – for example, if you continue to run workloads on the same machines after your plan expires.
Types of AWS Savings Plans
How are AWS Savings Plans different from Reserved Instances?
Companies use Reserved Instances because they offer significant reductions over On-Demand pricing. To use a Reserved Instance, you commit to a specific cloud capacity for a set time. In AWS, it’s either one or three years. AWS guarantees access to specific resources within a designated hosting region in certain scenarios.
Everything works the same way as standard instance purchases. Choose an instance type, size, platform, and area, and you’re ready! It’s similar to receiving a voucher that can be used for a discount at any moment throughout your chosen reservation period and is frequently shared across teams.
The larger your initial payment, the greater the discount.
How much can you save with Reserved Instances?
AWS claims EC2 Reserved Instances save up to 75% over On-Demand pricing.
But there is a catch. A Reserved Instance works on a “use it or lose it” premise. Every hour your instance is idle is lost (along with any monetary rewards). To get the most out of your Reserved Instance, you must maximize its use. This implies you need to know exactly what your team will require in advance.
Reserving capacity ahead of time to receive the instances you use frequently at a discount appears to be a wise decision. After all, you will continue to use these cloud resources. However, it comes with a price that many teams are unaware of.
Risks of Reserved Instances
1. High cost for changing requirements
When you commit to specific resources or usage levels, you expect your needs will not alter while the contract is active. However, a year’s commitment is an eternity in the cloud world. Not even IT giants with entire departments dedicated to cost efficiency can make accurate estimates.
2. Vendor lock-in
Entering into this form of contract with AWS exposes you to the danger of vendor lock-in. This means you’ll rely on AWS for one to three years. Using another provider’s products and services makes no sense if you have already paid for them. It would also incur significant switching expenses.
But what if AWS doesn’t have a solution for building something new after you sign the contract?
3. You lose all flexibility
Your requirements may alter in the future. When was the last time you had to deal with a requirement that remained unchanged for a year or two?
When faced with a new issue, your team may need to commit to even more or risk being stuck with unused capacity that you have already paid for. Either way, you’re on the losing end. You won’t be able to simply scale or configure multi-region/zone distribution.
4. Setting up AWS Reserved Instances is really complex
Choosing the best Reserved Instance to match your application’s workload requirements is difficult, to say the least. Keeping up with the latest AWS services is difficult enough, let alone selecting among roughly 770+ EC2 instance types. On top of that, you’ll need to assess your application’s requirements and workload patterns.
Manual processes are time-consuming and imprecise, making the whole endeavor extremely dangerous.
Using cloud bill analysis methods to calculate Reserved Instances or Savings Plan requirements is a bad idea. Their insights may lead you to reserve capacity based on your current instance, even if it’s the wrong type and size.
When you reserve the wrong instances, you risk overprovisioning your application and wasting a lot of money in the long run.
AWS Savings Plans and cloud waste
Whether you choose Compute Savings Plans or standard EC2 instance Savings Plans doesn’t matter. You must comprehend the complexities of savings plans and how they work.
You purchase a Savings Plan based on an hourly commitment. AWS utilizes Cost Explorer to automatically compute how that commitment will appear as a monthly fee on your AWS invoice.
Here’s an example:
The monthly On-Demand charge for a single EC2 instance is $2,000. The Savings Plan’s term and payment options result in a 30% savings. The recommended commitment will be around $1.92 per hour or $1,400 monthly.
Problems arise when the hourly resource utilization of the EC2 machine families you’re using begins to fluctuate.
If you overcommit resources, every hour your teams use less than the committed amount is a loss of value. If you don’t stick to your Savings Plan and keep going over it, you’ll end up paying for additional resources in the highest pricing tier, On-Demand.
When using the Savings Plan, you will still receive a discount on the cost of any EC2 instances that your teams provisioned. You won’t be able to take advantage of the Savings Plan discount if you overprovision an EC2 instance.
That is why it makes sense for many teams to purchase Savings Plans in increments. This helps prevent your organization from committing to more resources than it can use monthly.
Savings Plans are ideal for accounts where you can forecast minimal usage throughout the commitment period (one or three years). If you’re working with an application with an unknown baseline, it’s a smart idea to begin by gathering usage statistics and then gradually commit to Savings Plans.
Can you save money even if you’ve already committed to a Savings Plan?
Savings Plans can help you lower your AWS payment, but you are still in charge of infrastructure optimization.
That’s why rightsizing is such an important undertaking. If you manage a large cloud environment, you’ll need a system that automates rightsizing activities.
It takes time to determine which resources are used, which families own them, and which teams control them. Trying to make sense of all EC2 instances AWS offers (together with various pricing methods) is no easy feat. Assessing your inventory and utilization to determine which instances can be lowered may take many days or weeks.
Without your intervention, an AI-powered cloud cost optimization solution can instantly identify unused infrastructure components and degrade or terminate assets.
For example, Iterable used Cast AI automation to generate cost savings, and the capacity covered by Savings Plans allowed the team more flexibility.
“This is a good place to be in, but if you’re paying for certain reliability, you should use it. Now we can scale any feature, knowing that we have the ability to offset that growth with Spot Instances. And it’s cheaper growth – a third of the price. All the while, we keep our Savings Plans 100% percent utilized, not overutilized or underutilized. This gives us an extra lever that we didn’t have before to really control our spend.”
Jason Sanghi, Staff Software Engineer, SRE, Iterable
Discover how Iterable cut EKS costs by 60%
See how Iterable gained cost transparency, optimized spot usage, and slashed container expenses with Cast AI. Try it free or explore the full case study.
