7 Compute Optimization Strategies

 

This chapter covers

  • Understanding AWS compute services and their associated pricing models.
  • Assessing compute resource usage to identify cost-saving opportunities.
  • Implementing automation techniques to optimize resource utilization.
  • Implementing Kubernetes cost management and multi-cloud compute optimization.

The Optimize Phase in FinOps is all about achieving cost efficiency by aligning cloud resources with actual usage patterns. After building visibility into cloud costs during the Inform Phase, it’s time to take action to reduce waste, improve resource utilization, and ensure workloads run as cost-effectively as possible.

Compute resources often represent the largest portion of cloud spending, accounting for 60% to 70% of total cloud costs in many organizations, according to industry benchmarks. Without proper management, teams may overprovision instances, leave idle resources running, or rely on expensive pricing models that aren’t aligned with workload requirements. Addressing these inefficiencies not only lowers costs but also improves scalability and performance.

7.1 AWS Compute Services and Pricing Models

7.1.1 AWS Compute Services

7.1.2 Pricing Models

7.2 Assessing Compute Usage

7.2.1 Understanding Current Usage and Costs

7.2.2 Identifying Wasted Compute Resources

7.3 Implementing Automation for Cost Optimization

7.3.1 Scheduling On/Off Times for Instances

7.3.2 Implementing Spot Instances for CI/CD Pipelines

7.3.3 Lambda Memory Tuning

7.3.4 Rightsizing & Autoscaling Instances

7.3.5 Optimizing Kubernetes Compute Costs

7.3.6 The KPI and Modernization Dashboard

7.3.7 Managing Multi-Cloud Compute Resources

7.4 Summary