A healthcare technology company is architecting a medical diagnostics platform on Google Cloud. The system consists of two distinct components: 1) A stateless REST API backend that receives patient request payloads and returns quick diagnostic summaries with highly variable, bursty web traffic, requiring minimal operational management; 2) A batch processing pipeline that runs legacy, uncontainerized diagnostic algorithms relying on custom Linux kernel modules for 3 hours every night. Which TWO compute solutions should you recommend to fulfill these requirements with the lowest operational overhead and optimal cost efficiency? (Select TWO)
- Deploy the stateless REST API backend onto Cloud Run.Answer
- Deploy the batch processing pipeline onto Compute Engine Managed Instance Groups using Spot VMs.Answer
- CDeploy the stateless REST API backend onto a dedicated Google Kubernetes Engine (GKE) Standard cluster with fixed node pools.
- DDeploy the legacy batch processing pipeline onto Cloud Functions.
- EPurchase 3-year Committed Use Discounts (CUDs) for On-Demand Compute Engine instances to run the 3-hour nightly batch workload.
Answer
Select Cloud Run for the stateless REST API backend and Compute Engine Managed Instance Groups using Spot VMs for the legacy batch processing pipeline.
Cloud Run is ideal for stateless REST APIs with bursty traffic because it automatically scales containers and requires zero cluster maintenance. For legacy workloads requiring custom kernel modules and running for short nightly windows, Compute Engine VMs provide the required OS-level access, while Spot VMs and Managed Instance Groups ensure maximum cost efficiency.
Step-by-Step Solution
Key Concept
Selecting appropriate compute platforms based on containerization status, kernel/OS customization needs, scaling patterns, and management overhead constraints.