Examkin

Question

Difficulty: HardSelecting and Designing Compute and Application Platforms

An IoT smart city energy utility enterprise is designing its cloud solution architecture on Google Cloud to handle device management and telemetry processing. The architecture comprises two distinct workloads:

1. A stateless HTTP ingestion endpoint receiving intermittent spikes of JSON payloads from smart meters. The service must scale rapidly during peak hours, scale down to zero instances when idle to reduce costs, and require minimal infrastructure management.
2. A specialized network protocol parser running a long-lived daemon that requires custom OS kernel sysctl modifications for low-level socket buffer tuning and high-speed local NVMe scratch disk access.

Which TWO architectural choices should you select to meet these requirements with optimal operational efficiency? (Select TWO.)

  1. Deploy the stateless HTTP ingestion endpoint on Cloud Run with container concurrency and auto-scaling configured.Answer
  2. Deploy the specialized network protocol parser on Google Kubernetes Engine (GKE) nodes configured with local SSD storage and custom node system configurations for sysctl tuning.Answer
  3. C
    Deploy the stateless HTTP ingestion endpoint on a dedicated Google Kubernetes Engine (GKE) Standard cluster configured with Horizontal Pod Autoscaler.
  4. D
    Deploy the specialized network protocol parser on Cloud Run by packaging the custom sysctl parameters into the container build manifest.
  5. E
    Deploy the stateless HTTP ingestion endpoint on a Compute Engine Managed Instance Group (MIG) behind an External HTTP(S) Load Balancer.

Answer

Select the stateless HTTP ingestion endpoint deployment on Cloud Run, and the specialized network protocol parser deployment on Google Kubernetes Engine (GKE) with custom node configuration and Local SSDs.
Cloud Run is the optimal compute platform for stateless HTTP services with variable traffic because it automatically scales down to zero when idle and eliminates server maintenance overhead. For workloads requiring custom host OS kernel modifications (sysctl parameters) and direct hardware attachments like Local SSDs, Google Kubernetes Engine (GKE) provides the necessary node configuration flexibility while remaining a managed orchestration platform.

Step-by-Step Solution

1
Analyze workload 1 requirements
Workload 1 requires stateless HTTP processing, rapid autoscaling to zero, and minimal operational overhead.
Cloud Run fits fully managed stateless HTTP container requirements perfectly without incurring node baseline costs or cluster administration duties.
2
Analyze workload 2 requirements
Workload 2 requires host kernel parameter tuning (sysctl) and high-speed local NVMe disk access.
GKE allows custom sysctl settings via system configuration and supports Local SSD node pools, whereas fully managed serverless environments restrict host kernel modification and direct hardware access.

Key Concept

Selecting GCP Compute Platforms based on workload statefulness, scaling characteristics, kernel tuning demands, and operational overhead limits.
Rate this question

Topics

Selecting and Designing Compute and Application Platforms