top of page
  • Writer's pictureShannon

Reserved Instances (RIs) vs. Savings Plans (SPs): A Comparison between Azure and AWS

This post required a revision based upon how I originally positioned AWS Reserved Instances. I'm still learning AWS, so I had some

During the day, I spend so much time talking about what a Reserved Instance is vs. a Savings Plan. I figured a blog post seems like the next logical step getting content out to the broader community.

In the ever-evolving landscape of cloud computing, cost optimization remains a top priority for...well...everyone. Both Microsoft Azure and Amazon Web Services (AWS) offer powerful tools to help you save on your cloud expenses. Let’s next dive into the details of Reserved Instances (RIs) and Savings Plans (SP), comparing their features, use cases, and benefits.

1. What Are Reserved Instances (RIs)?

  • Azure RIs: When you purchase an RI, you commit to using a specific virtual machine (VM) type in a particular Azure region for either 1 or 3 years. RIs provide substantial savings – up to 75% off On-Demand pricing.

  • AWS RIs: AWS RIs allow to save up to 75% on on-demand instances by purchasing then in advanced for a fixed term (either one or three years).

2. Taking it One Step Deeper: What is the Difference Between AWS RIs and On-Demand Capacity Reservations?

Trust me folks...there is a distinct difference here and I only recently learned the difference by reading this article.

  1. AWS Reserved Instances:

  • Reservation Model: With RIs, you “book” a specific amount of computing power and pay upfront.

  • Payment Options:

  • Standard RIs: Pay a lower per-second or per-hour rate for the reserved capacity.

  • Convertible RIs: Offers flexibility to change instance types or operating systems during the reservation term.

  • Scheduled RIs: Reserve capacity for specific time windows (e.g., daily or weekly).

  • Commitment: Requires a fixed-term commitment.

  • Capacity Reservation: Optionally provides a capacity reservation for instances.

  • Billing: Upfront payment for the reserved capacity.

  • Usage: Ideal for predictable workloads with long-term requirements.

  1. On-Demand Capacity Reservations:

  • Purpose: Capacity Reservations allow you to reserve compute capacity for Amazon EC2 instances in a specific Availability Zone for any duration.

  • Reservation Model: No commitment required; create and cancel reservations as needed.

  • Billing: Pay for compute capacity by the second with no long-term commitments.

  • Usage: Suitable for scenarios where you need capacity assurance without committing to a fixed term.

3. What Are Savings Plans (SPs)?

  • Azure SPs: Azure SPs are flexible pricing models that let you commit to spending a fixed hourly amount collectively on compute services. These apply to various usage types, including EC2, Fargate, and Lambda.

  • AWS SPs: AWS SPs provide savings of up to 72% on compute usage. Unlike RIs, SPs cover any instance family, size, OS, tenancy, or AWS region. They also apply to Fargate and Lambda usage.

4. Key Differences: RIs vs. SPs (Broad Themes)

  • Instance Coverage:

  • RIs: Specific reservations for virtual machines.

  • SPs: Cover any instance family, size, and region.

  • Flexibility:

  • RIs: Less flexible; tied to specific VM types.

  • SPs: Accommodate changes in compute needs without modifications.

  • Applicability:

  • RIs: Only applicable to virtual machines.

  • SPs: Apply to everything at the compute layer (web apps, Fargate, AKS, EKS, etc.)

  • Discount Model:

  • RIs: Discount against On-Demand pricing based on committed utilization.

  • SPs: Discount based on committed compute spend.

5. When to Use Each?

  • Choose RIs:

  • For stable workloads with minimal changes to VM types or regions.

  • When maximum savings are crucial.

  • Choose SPs:

  • For dynamic workloads that require flexibility.

  • When you need to run different-sized VMs or frequently change datacenter regions.

6. Conclusion

  • Azure: RIs for stability, SPs for flexibility.

  • AWS: RIs and SPs complement each other; use them wisely based on your workload patterns.

Remember, cloud cost optimization isn’t a one-size-fits-all approach (gosh, I wish it'd make my daily life a little easier at times). Evaluate specific needs and make informed decisions to maximize savings while meeting business requirements.

Happy cloud cost management! 🌟

Helpful References:

Recent Posts

See All


bottom of page