Design Cost-Optimized Compute Solutions
Compute is an area in AWS that can lead to spiraling costs if not optimized correctly. Several strategies can be employed to help you optimize your compute spend, such as selecting the best pricing model for your workload, selecting the right compute platform for your workload, and leveraging regional savings.
AWS Compute Service Pricing Models
In Chapter 4, Compute, you learned about the different instance purchase models:
- On-Demand: Standard instance purchase option where you pay for the instance family type, length of operation, and OS type on a per-hour or per-second basis, depending on the OS selected.
- Reserved Instances: Reserve instance capacity for 1 or 3 years, paying either all upfront, partial upfront, or none upfront.
- Spot: Bid for spare compute capacity.
- Savings Plans: Reserve compute capacity for 1 or 3 years. Includes EC2 instances, Lambda, Fargate, and SageMaker.
- Dedicated Hosts/Instances: Reserve...
