DEV Community

Cover image for AWS Under the Hood - Day 8 - Unlock Hidden Savings in AWS with Savings Plans
Prashant Lakhera
Prashant Lakhera

Posted on

AWS Under the Hood - Day 8 - Unlock Hidden Savings in AWS with Savings Plans

When discussing cost savings in AWS, the conversation often revolves around:

  • On-Demand
  • Reserved Instances
  • Spot Instances

However, one aspect that is less widely known is Savings Plans.
Savings Plans offer a flexible pricing model that can help you save up to 72% on your usage of EC2, Fargate, and Lambda services. There are two types of Savings Plans:

  • Compute Savings Plans
  • EC2 Instance Savings Plans

What are Savings Plans?

Savings Plans are a flexible pricing model offered by Amazon Web Services (AWS) that provides significant savings on your AWS usage (66% up to 72%) in exchange for a commitment to a consistent amount of usage (measured in $/hour) for a 1 or 3 year period. There are mainly two types of Savings Plans:

  • 1: Compute Savings Plans
    These plans offer savings for any usage regardless of region, instance family, operating system, or tenancy, as long as it's EC2, Fargate, or Lambda.
    They are highly flexible and are ideal for users with fluctuating workloads, applications that span multiple regions, or applications that might shift from one instance family or operating system to another.

  • 2: EC2 Instance Savings Plans
    These provide savings focused specifically on EC2 instances in a single region, regardless of the OS or tenancy. However, these plans are limited to the instance family.
    For example, if you commit to a specific amount of usage on C5 instances in the North Virginia region, you'll receive discounted rates on any C5 usage in that region

  • 3: Comparison with Reserved Instances (RIs)
    Reserved Instances also provide a way to save on EC2 instances by committing to a specific instance configuration (including instance type and region) for either a 1-year or 3-year term.

There are some key differences between RIs and Savings Plans:

  • 1: Flexibility
    Savings Plans: These are more flexible, as they allow shifting usage across different instance types, regions, or even AWS services (within Compute Savings Plans).
    Reserved Instances: Less flexible, committing you to specific instance types and regions. Any changes in usage patterns not covered by the RI specifications won't benefit from the RI discounts.

  • 2: Discount Levels
    Both offer similar discounts, but how those discounts apply can vary depending on the commitment level and the options chosen (like payment options).

  • 3: Payment Options
    Savings Plans: Offers three payment options - All upfront, partial upfront, or no upfront.
    Reserved Instances: Offers similar payment options but also includes convertible RIs which provide more flexibility in changing the attributes of the RI in exchange for a slightly lower discount compared to standard RIs.

  • 4: Applicability
    Savings Plans: Apply broadly across your bill, automatically adjusting to usage patterns.
    Reserved Instances: These need to be managed and tracked more carefully to ensure they are being fully utilized, or else the savings potential is lost.

How it works Under the Hood
When you commit to a Savings Plan or RI, AWS effectively locks in a discount rate for the agreed-upon usage. This rate is applied to your bill for the duration of the term. The actual mechanics involve monitoring your usage every hour:
If your usage exceeds the committed amount (in the case of Savings Plans), standard on-demand rates apply to the excess.
If your usage falls below the committed amount, you still pay for the committed usage level.

This model benefits AWS by providing predictable cash flow and utilization patterns, which allows them to optimize their data center operations and capacity planning. For users, it provides a significant discount in exchange for predictable billing and some level of commitment.
Savings Plans offer greater flexibility and ease of use, particularly for diverse or variable workloads. At the same time, RIs might be preferred for very stable and predictable usage patterns that are unlikely to change.

🛑 Despite the commitment to a Saving plan, the actual allocation of resources (like instances) occurs on an on-demand basis. You are free to launch or terminate instances as your needs change, and the discount will apply to whatever eligible usage occurs each hour up to the limit of your Savings Plan

📚 If you're interested in more in-depth explanation of these topics, please check out my new book "Cracking the DevOps Interview"

📚 To learn more about AWS, check out my book "AWS for System Administrators"

Pic reference:

Top comments (0)