A financial analytics enterprise processes a continuous low-volume stream of transactional data that requires a regional relational database with low administrative overhead. Additionally, the platform executes high-throughput batch analysis jobs for 48 hours once per month, requiring immense compute capacity that remains idle for the rest of the month. The enterprise wants to optimize Google Cloud infrastructure costs while maintaining high availability and scaling dynamically during peak processing. Which TWO architectural decisions should the cloud architect implement to satisfy these business requirements? (Select TWO.)
- AProvision a multi-region Cloud Spanner instance with high node count to host the transactional database.
- Deploy the transactional database on Cloud SQL for PostgreSQL with automatic storage increase enabled.Answer
- Execute the monthly 48-hour batch analysis jobs using Compute Engine Spot VMs in a Managed Instance Group with checkpointing.Answer
- DPurchase 3-year resource-based Committed Use Discounts (CUDs) covering the peak compute instance count required for the monthly batch analysis jobs.
Answer
Deploy the transactional database on Cloud SQL for PostgreSQL with automatic storage scaling, and execute the monthly 48-hour batch analysis workloads on Compute Engine Spot VMs with checkpointing.
Selecting Cloud SQL for PostgreSQL satisfies the requirement for a low-overhead regional relational database at minimal baseline cost. Utilizing Compute Engine Spot VMs with job checkpointing for the 48-hour monthly batch workload achieves dramatic cost reductions compared to standard provisioned instances while supporting fault tolerance.
Step-by-Step Solution
Key Concept
Selecting cost-effective compute and storage architectures matching workload intermittency and regional constraints.