AWS Reserved Instances (RIs) come in two primary types: Standard Reserved Instances and Convertible Reserved Instances. Both offer significant cost savings compared to On-Demand pricing in exchange for committing to a 1- or 3-year term, but they differ in terms of flexibility, usage, and savings.
1. Standard Reserved Instances (RIs)
Key Characteristics:
- Commitment: When you purchase a Standard Reserved Instance, you commit to using a specific instance type, region, availability zone, and operating system for the entire term (1 or 3 years).
- Pricing: Standard RIs generally offer the highest savings (up to 75% compared to On-Demand prices). The savings depend on the payment plan you choose (All Upfront, Partial Upfront, or No Upfront).
- Flexibility: The flexibility is limited—you can’t change the instance type, region, or operating system during the term of the reservation. However, you can modify some details like the Availability Zone (within the same region) or instance size (within the same instance family) in certain cases.
Use Case:
- Best for workloads with stable, predictable usage that won’t change frequently. Examples include applications with fixed compute requirements, long-running databases, or other steady-state workloads.
Pros:
- Highest savings.
- Simple to manage, since it locks in your instance usage.
- Can be ideal for long-term, steady-state applications.
Cons:
- Less flexibility—once the RI is purchased, you are locked into the specific configuration (instance type, region, etc.).
- Harder to adapt to changing workloads or different instance types.
2. Convertible Reserved Instances (RIs)
Key Characteristics:
- Commitment: Convertible RIs offer more flexibility than Standard RIs. You commit to a 1- or 3-year term, but you can change the instance type, region, operating system, or tenancy during the term of the reservation.
- Pricing: Convertible RIs provide lower savings compared to Standard RIs (typically up to 54% savings compared to On-Demand pricing), but the tradeoff is the ability to make modifications.
- Flexibility: The major advantage of Convertible RIs is the flexibility to exchange your existing Reserved Instance for a new one with different attributes, such as:
- Instance type (within the same family)
- Region
- Operating system
- Tenancy (shared or dedicated)
Use Case:
- Best for workloads with changing or evolving requirements, such as growing applications, experimental environments, or workloads that may require scaling or shifting resources (e.g., machine learning, development environments).
Pros:
- Greater flexibility—easily adjust instance types, regions, or operating systems as your needs evolve.
- Can swap one Convertible RI for another during the term without penalty.
- More suitable for dynamic workloads.
Cons:
- Lower savings compared to Standard RIs (though still a discount over On-Demand).
- Slightly more complex than Standard RIs due to flexibility.
Comparison Table:
Feature | Standard Reserved Instances | Convertible Reserved Instances |
---|---|---|
Commitment | 1 or 3 years | 1 or 3 years |
Flexibility | Limited (instance family, AZ, OS) | High (can change instance type, region, OS, tenancy) |
Savings | Up to 75% savings compared to On-Demand | Up to 54% savings compared to On-Demand |
Instance Type Modifications | Limited (same instance family) | Full flexibility (instance family, size, region, OS) |
Best For | Stable, predictable workloads | Dynamic, evolving workloads |
Examples | Long-term databases, steady-state apps | Test environments, applications with scaling needs |
Payment Options | All Upfront, Partial Upfront, No Upfront | All Upfront, Partial Upfront, No Upfront |
Summary:
- Standard Reserved Instances are best for predictable, steady workloads where you know you’ll be using the same instance type, region, and operating system for a long time. They offer the greatest savings but come with limited flexibility.
- Convertible Reserved Instances are ideal if your needs are likely to change over time. They offer lesser savings than Standard RIs but provide much more flexibility, allowing you to adjust your instances to meet evolving requirements without losing your investment.
If you are certain that your compute needs won’t change, Standard RIs might be the best choice for maximizing cost savings. However, if your needs are more flexible or you expect to scale and adjust over time, Convertible RIs would be a better fit.
Summary of What You Can Change:
Feature | Standard RIs | Convertible RIs | Zonal RIs |
---|---|---|---|
Region | No | Yes | No |
Availability Zone | Yes (within the same region) | Yes (within the same region) | Yes (within the same region) |
Instance Size | Yes (within the same instance family) | Yes (within the same instance family) | Yes (within the same instance family) |
Instance Family | No | Yes | No |
Operating System | No | Yes | No |
Tenancy | Yes (Shared/Dedicated) | Yes (Shared/Dedicated) | Yes (Shared/Dedicated) |
Term | No (fixed once purchased) | No (fixed once purchased) | No (fixed once purchased) |
Reserve compute capacity with EC2 On-Demand Capacity Reservations
Amazon EC2 Capacity Reservations allow you to reserve compute capacity for your Amazon EC2 instances in a specific Availability Zone for any duration. If you have strict capacity requirements for current or future business-critical workloads that require a certain level of long or short-term capacity assurance, we recommend that you create a Capacity Reservation to help ensure that you always have access to Amazon EC2 capacity when you need it, for as long as you need it.
You can create a Capacity Reservation at any time, and you can choose when it starts. You can request a Capacity Reservation for immediate use or you can request a Capacity Reservation for a future date.
- If you request a Capacity Reservation for immediate use, the Capacity Reservation becomes available for use immediately and there is no term commitment. You can modify the Capacity Reservation at any time, and you can cancel it at any time to release the reserved capacity and to stop incurring changes.
- If you request a future-dated Capacity Reservation, you specify the future date at which you need the Capacity Reservation to become available for use. You must also specify a commitment duration for which you commit to keeping the requested capacity in your account after the specified date. At the requested date and time, the Capacity Reservation becomes available for use and the commitment duration starts. During the commitment duration, you can’t decrease the instance count or commitment duration below your initial commitment, or cancel the Capacity Reservation. After the commitment duration elapses, you can modify the Capacity Reservation in any way or cancel it if you no longer need it.
Capacity Reservations can only be used by instances that match their attributes. By default, Capacity Reservations automatically match new instances and running instances that have matching attributes (instance type, platform, Availability Zone, and tenancy). This means that any instance with matching attributes automatically runs in the Capacity Reservation. However, you can also target a Capacity Reservation for specific workloads. This allows you to explicitly control which instances are allowed to run in that reserved capacity. You can also specify that instances will only run in a Capacity Reservation or Capacity Reservation resource group.