02-26-2025 / Cost Optimization Strategies / 6 mins.
What Are AWS Reserved Instances?
Reserved Instances (RIs) offer discounted hourly rates and an optional capacity reservation for Amazon EC2 and RDS services. By committing to a specific instance type and region for a predefined period, users can reduce their compute costs by up to 75% compared to On-Demand instance pricing.
This reservation model is ideal for workloads with consistent usage, such as a control plane component that is always running regardless of the number of worker nodes. Reserving an instance involves choosing a term (one or three years), an instance type, and payment options. These decisions determine the required upfront payment and overall savings.
Benefits of Reserved Instances
Reserved Instances are a cornerstone of FinOps strategies on AWS, offering predictable costs for long-term workloads, substantial savings over On-Demand rates, and the ability to reserve capacity in a specific availability zone. These benefits are particularly valuable for companies implementing FinOps, as they allow predictable costs to align with optimized budgets and resource planning.
Beyond cost savings, RIs improve stability in application deployments, especially in scenarios where capacity is limited. Organizations can ensure that their critical applications run without interruptions or capacity risks. Predictable costs and savings facilitate better budget management and financial planning — key components of an effective FinOps strategy on AWS.
On-Demand Instances vs. Reserved Instances vs. Convertible Reserved Instances vs. Savings Plans: What Is the Difference?
- On-Demand Instances: Allow organizations to pay for compute capacity by the hour or second (depending on the instance type) with no long-term commitments. They offer the flexibility to scale up or down as needed. However, they are the most expensive way to run instances.
- Reserved Instances (RIs): Require a commitment to use a specific instance type in a chosen region for one or three years, in exchange for a lower rate. While they offer less flexibility to change or cancel service, the predictable pricing and substantial savings are valuable for many use cases.
- Convertible Reserved Instances: Offer more flexibility than standard RIs by allowing you to change instance attributes (family, operating system, tenancy type, and payment option) during the commitment period. Convertible RIs carry smaller discounts than standard ones, but are useful for applications that may require changes over time.
- Savings Plans: Offer the flexibility of On-Demand Instances along with significant savings similar to RIs. There are two types:
- Compute Savings Plans: Automatically apply to any instance usage across AWS, regardless of instance family, region, operating system, or tenancy.
- EC2 Instance Savings Plans: Apply to specific instance families in chosen regions, offering slightly higher discounts.
How Do Reserved Instances Work?
The typical process for purchasing RIs is as follows:
- Users select the instance type, operating system, and region. If needed, they can also specify an availability zone to reserve capacity within that zone, which is useful for ensuring resource availability in high-demand areas.
- Once purchased, RIs are automatically applied to running instances that match the reservation attributes, reducing their hourly rate.
- If usage exceeds the reserved capacity, additional instances are billed at the On-Demand rate.
- When the RI term ends, instances continue running but are automatically switched to On-Demand pricing.
Types of AWS Reserved Instances
- Reserved Instances - Amazon EC2: Allow users to reserve an instance type and operating system in a specific region and availability zone. This is useful for planning application infrastructures that require specific configurations and reliable performance. EC2 RIs are common in production environments where a consistent workload is expected.
- Reserved Instances - Amazon RDS: Allow users to secure discounted rates for their database instances by committing for one or three years. This model is especially advantageous for applications with consistent and predictable long-term database workloads.
- Users select the database engine (MySQL, PostgreSQL, Oracle, etc.), instance class, and region when purchasing RDS Reserved Instances.
- They can choose between standard or Multi-AZ deployments. Multi-AZ deployments offer high availability and failover support, ideal for mission-critical applications.
Tips
- Analyze historical data before purchasing Reserved Instances:Before committing to Reserved Instances, use Cost Explorer to assess your past usage trends over at least 6 months. This will help ensure your RI purchases are aligned with long-term resource requirements and prevent over-reservation.
- Use a combined Reserved Instance and On-Demand Instance strategy:Rather than going all-in on RIs, balance your infrastructure with a mix of Reserved and On-Demand Instances. This gives you the flexibility to scale while taking advantage of savings for predictable workloads.
- Opt for Convertible RIs in evolving environments:If you anticipate changes to your infrastructure over the next 1–3 years (e.g., shifts in workloads or instance types), Convertible RIs offer the flexibility to adjust without sacrificing too much in terms of savings.
- Leverage the RI Marketplace to fine-tune commitments:If you have over-committed to a specific RI type, consider selling it on the "Reserved Instance Marketplace". This secondary marketplace allows you to recover part of the costs and adapt your reservations to the changing needs of your infrastructure.
- Continuously monitor Reserved Instance utilization:Use AWS Trusted Advisor and Cost Explorer to regularly check whether you are fully utilizing your RIs. Underutilized RIs represent wasted capital, and reassigning workloads to use them effectively can maximize your return on investment (ROI).
How to Take Advantage of Reserved Instances with Frust
With Frust you can obtain the same financial benefits as AWS Reserved Instances without having to commit to 12- or 36-month contracts, allowing you to maintain flexibility in managing your AWS infrastructure costs.
