Professional Cloud ArchitectDesign and plan a cloud solution architectureEasy
A global media company is planning to migrate its large-scale video processing workflows to Google Cloud. These workflows are highly parallelizable, fault-tolerant, and can be interrupted and restarted without significant impact. They need a cost-effective solution for processing petabytes of data daily. Which Google Cloud compute option should they choose?
- ABatch
- BGoogle Kubernetes Engine (GKE) Autopilot
- CCompute Engine custom machine types
- DCloud Functions
Show answer & explanationAnswer & explanation
Correct answer: A. Batch
Batch is designed for large-scale, fault-tolerant, and parallelizable batch processing workloads, offering cost-effectiveness through optimized resource utilization. Its managed nature simplifies job execution and resource management for such tasks.
Why the other options are wrong
- B. GKE Autopilot is excellent for containerized applications but might be overkill and less cost-optimized for purely batch, interruptible workloads compared to a dedicated batch service.
- C. Custom machine types offer flexibility but don't inherently provide the managed batch processing capabilities or the same cost optimization for interruptible, large-scale jobs.
- D. Cloud Functions are suitable for event-driven, short-running, and stateless functions, not for petabyte-scale, long-running video processing workflows.
Google Cloud Batch
A fully managed service for running large-scale batch jobs on Google Cloud, optimizing resource allocation and job execution.
- Ideal for HPC, data processing, and scientific simulations.
- Supports various machine types, including custom and GPUs.
- Integrates with other Google Cloud services like Cloud Storage and Pub/Sub.
Memory trick: Batch processes big jobs, breaking them into bite-sized bits.