Examkin

Question

Difficulty: HardSelecting and Designing Compute and Application Platforms

An energy Grid IoT analytics company is modernizing its telemetry processing platform on Google Cloud. The system must support two distinct workloads with separate operational characteristics:

1. Event Ingestion Microservice: An HTTP REST service that receives telemetry webhooks from smart meters. Traffic fluctuates unpredictable throughout the day and drops to zero overnight in specific regions. The service must scale rapidly, minimize management overhead, and scale down to zero when idle.
2. Batch Spatial Simulation Engine: An intensive data pipeline requiring custom Linux kernel parameters (`sysctl` network tuning) and direct `/dev/shm` shared memory allocations. The workload processes batched data chunks periodically, tolerates unexpected instance terminations, and must run at the lowest possible infrastructure cost.

Which TWO compute architecture decisions should the Cloud Architect recommend to satisfy these requirements? (Select TWO.)

  1. Deploy the Event Ingestion Microservice on Cloud Run.Answer
  2. B
    Deploy the Event Ingestion Microservice to a dedicated multi-zone Google Kubernetes Engine (GKE) cluster.
  3. Deploy the Batch Spatial Simulation Engine on Compute Engine Managed Instance Groups (MIGs) utilizing Spot VMs.Answer
  4. D
    Purchase 3-year Committed Use Discounts (CUDs) for fixed-size Compute Engine VM instances for both workloads.
  5. E
    Deploy the Batch Spatial Simulation Engine on Cloud Functions (2nd gen) with default container settings.

Answer

The optimal solution uses Cloud Run for the Event Ingestion Microservice and Compute Engine Managed Instance Groups with Spot VMs for the Batch Spatial Simulation Engine.
Cloud Run provides serverless HTTP execution that scales to zero for fluctuating REST webhook workloads without infrastructure management. Compute Engine Managed Instance Groups running Spot VMs provide the required OS-level access for custom kernel sysctl parameters and shared memory while maximizing cost efficiency for fault-tolerant batch jobs.

Step-by-Step Solution

1
Analyze the Event Ingestion Microservice requirements
Identified stateless HTTP REST protocol, rapid auto-scaling demands, requirement to scale down to zero idle instances, and desire for low management overhead.
Cloud Run is purpose-built for stateless HTTP container workloads, managing server provisioning and scaling to zero automatically.
2
Analyze the Batch Spatial Simulation Engine requirements
Identified low-level OS configuration needs (sysctl, shared memory), fault-tolerant processing, and cost minimization goals.
Compute Engine VMs permit custom Linux kernel modifications and shared memory configuration. Spot VMs offer up to 60-91% savings for fault-tolerant workloads.
3
Evaluate alternative options for misallocations
GKE adds unnecessary control plane management overhead for simple microservices, Cloud Functions lacks host kernel parameter access, and CUDs misalign with variable/preemptible workload profiles.
Architecting GCP compute solutions requires matching technical constraints and cost profiles to the correct execution platform.

Key Concept

Selecting GCP Compute Platforms based on operational overhead, scaling behavior, host OS customization needs, and cost optimization.
Rate this question

Topics

Selecting and Designing Compute and Application Platforms