An organization is planning its Google Cloud infrastructure for two distinct applications. Application 1 is a high-throughput video transcode batch processing pipeline that processes stateless chunks in parallel and can handle unexpected instance terminations without data loss. Application 2 is a core web service operating continuously 24 hours a day, 7 days a week, with predictable baseline resource requirements. Which TWO machine planning and pricing strategies should you implement to optimize overall compute costs while maintaining performance requirements?
- Provision Spot Compute Engine VMs for the stateless batch processing workers in Application 1.Answer
- Purchase Committed Use Discounts (CUDs) for the steady-state Standard VMs running Application 2.Answer
- CDeploy Application 2 on Spot VMs to minimize the baseline operational cost of continuous web services.
- DMigrate Application 1 batch transcode workers to Cloud Functions to avoid managing Compute Engine virtual machine types.
Answer
Select Spot Compute Engine VMs for the fault-tolerant batch processing workload, and apply Committed Use Discounts (CUDs) to Standard VMs for the continuous 24/7 web application.
The combination of Spot Compute Engine VMs for fault-tolerant batch processing and Committed Use Discounts (CUDs) for continuous 24/7 web services aligns directly with GCP cost-optimization best practices. Spot VMs drastically reduce compute costs for jobs that easily resume on preemption, while CUDs provide predictable price reductions for uninterrupted baseline infrastructure.
Step-by-Step Solution
Key Concept
Selecting appropriate Compute Engine purchasing models (Spot VMs vs. Committed Use Discounts) based on workload fault-tolerance and baseline usage predictability.