Major cloud providers offer customers ways to realize significant cost savings by committing to a specific level of consumption. Amazon Web Services calls this concept Amazon Reserved Instances (RIs) and provides them as an alternative to standard On-Demand instance pricing. RIs are a reservation of resources and capacity—not a physical instance type—and manifest as a billing discount or coupon applied against On-Demand instance use in an AWS account. RIs exist for EC2, databases, and CDNs.
Unfortunately, because each RI type is somewhat unique with its own break even point, an optimized strategy for Reserved Instance purchasing is critical. Without a real strategy (one that best comes from machine-generated analysis) any savings made by committing up front can easily be erased in practice.