Pricing models (for example, Reserved Instances, Savings Plans)

Can you explain the key differences between Reserved Instances and Savings Plans in AWS pricing models?

Reserved Instances (RIs) provide a significant discount compared to On-Demand pricing in exchange for a commitment to use a specific instance type, in a specific region, for a one or three-year term. Savings Plans, on the other hand, offer more flexibility; they provide cost savings that apply to any instance usage regardless of region or instance family, in exchange for a commitment to a consistent amount of usage (measured in $/hour) for one or three years. Unlike RIs, Savings Plans automatically apply the cost savings to any usage that matches the committed level.

When would you recommend using Reserved Instances over Savings Plans for an AWS environment?

Reserved Instances would be recommended when you have predictable usage patterns for specific instance types and when you can commit to using the same instance types in the same region. RIs are particularly well-suited for workloads with constant performance requirements, where the flexibility offered by Savings Plans is not needed.

What are Convertible Reserved Instances and how are they different from Standard Reserved Instances?

Convertible Reserved Instances give you the flexibility to change the attributes of the RI as long as the exchange results in equal or more compute power. This includes changes to instance families, OS type, and tenancy, offering greater adaptability if your needs change. Standard Reserved Instances don’t offer this flexibility but tend to offer higher discounts for the fixed attributes set at purchase.

What factors should a company consider when deciding between 1-year and 3-year commitment terms for AWS Savings Plans or Reserved Instances?

Companies should consider their long-term usage patterns, budget constraints, and potential changes in their computing needs. A 3-year term typically offers a higher discount rate but requires a longer-term commitment, which can be riskier if project needs are volatile. A 1-year term offers less discount but reduces the commitment period, providing a better balance between cost-saving and flexibility for many use cases.

How do the payment options for AWS Reserved Instances affect the discount received, and what are the different payment options available?

The discount for Reserved Instances varies with the payment option selected, with three main choices: All Upfront, Partial Upfront, and No Upfront. All Upfront provides the highest discount because you pay for the entire RI term at the start. Partial Upfront offers a lower discount, with part of the cost paid upfront and the rest billed monthly. No Upfront has the smallest discount but does not require any initial payment. Over time, the average cost will be lowest with All Upfront and highest with No Upfront.

What does it mean for an AWS Reserved Instance to be ‘zonal’ versus ‘regional,’ and how does this affect usage?

A ‘zonal’ Reserved Instance is tied to a specific Availability Zone, providing both a capacity reservation and a pricing discount. This can be important for applications requiring a guarantee of instance availability in a specified zone. A ‘regional’ Reserved Instance, on the other hand, provides pricing discounts across an entire region without capacity reservation, offering broader flexibility for where workloads can run.

Can you describe how AWS Savings Plans adapt to changes in instance usage?

AWS Savings Plans offer significant cost savings similar to Reserved Instances but with added flexibility. They automatically apply savings to any EC2 instance usage that matches the commitment, regardless of instance family, size, AZ, region, operating system, or tenancy. If usage drops below the commitment, the full benefit may not be realized, but if usage increases, the excess usage is charged at On-Demand rates.

What is an EC2 Instance Savings Plan and how does it differ from a Compute Savings Plan?

An EC2 Instance Savings Plan applies to a specific instance family within a region and is applied to individual instance usage regardless of the OS, tenancy, or AZ, as long as the instance family matches. A Compute Savings Plan, however, offers even greater flexibility, applying to any EC2 and Fargate usage across all regions and instance families, as long as the total hourly commitment is met.

How can you track and manage your Reserved Instances to ensure they are being utilized efficiently?

You can track and manage Reserved Instances through the AWS Management Console, using the Reserved Instance Utilization and Coverage reports within AWS Cost Explorer. Additionally, AWS provides tools such as Cost Allocation Tags and the AWS Budgets tool to monitor RI usage and expenses, allowing for adjustments as needed to maximize ROI and efficiency.

Is it possible to sell unused Reserved Instances, and if so, how does the AWS Reserved Instance Marketplace work?

Yes, it is possible to sell unused Standard Reserved Instances on the AWS Reserved Instance Marketplace, giving customers the flexibility to offload commitments if their needs change. Sellers can list their RIs at their desired price, and buyers can search for and purchase them, typically at a lower price point than directly from AWS. However, some restrictions apply, such as the RI must be held for at least 30 days and cannot be a Convertible RI.

What happens if you choose a Savings Plan or Reserved Instance and later need to upgrade to a different instance type or scale your infrastructure significantly?

If you are using Convertible Reserved Instances, you can exchange your existing RIs for others that better match your new requirements, with some limitations. With Savings Plans, the discount automatically applies to your usage as long as it meets the commitment level, offering natural flexibility for upgrades or scaling. However, if your upgrade leads to usage that exceeds your commitment, additional usage will be billed at the applicable On-Demand rate.