Amazon Reserved Instance Calculator

I’ve spent more time than I care to admit in spreadsheets trying to understand how to optimize the usage of Amazon reserved instances. So last weekend I set out on a mission: to develop a free service to assist AWS customers like myself in understanding how best to utilize reserved instances. After about 18 hours of hacking and a substantial amount of green tea, I produced a crude but functional version of what I call the Reserved Instance Analyzer.

What Is a Reserved Instance?

Reserved instances allow you to make a one-time payment to your cloud provider in return for a discounted hourly charge. Amazon and IBM are currently the only two public cloud vendors with reservation pricing, each with very different incentives and models. The cloud vendor rationale for reservations is simple: by getting customers to commit to the usage of specific infrastructure, they can reduce their operating costs and pass these savings on to their customers.

Amazon Reserved Instances

Amazon provides three types of reserved instance: light, medium and heavy. Each type offers a different prepayment and hourly cost for usage. Heavy reserved instances work slightly differently than medium and light in that you will be charged each month for 100% utilization regardless of actual usage. There is also an option with heavy to prepay the hourly utilization upfront for additional discounts (available only for volume accounts). Reservations are tied to availability zones, and can be made for either one of three years.

Confused yet? ;)

Reserved Instance Analyzer

The goal with the Reserved Instance Analyzer was to keep the process of analyzing reservations as simple as possible. Using the service involves three steps: 1) provide the accounts you want analyzed, 2) supply the projected utilization and usage of your instances, and then 3) receive the recommendation.

I have much left to do in order to make this a complete utility, but made it available in hope others can get value. Let me know what you think.