A digital media publishing company receives irregular batches of user-generated promotional video uploads. The raw video files must be processed and converted into multiple streaming formats. The processing workloads are stateless, highly parallelizable, and tolerant of temporary interruptions or delayed retries. The company wants to minimize compute costs for this batch processing pipeline while ensuring that raw uploads are automatically transitioned to lower-cost storage after 30 days. Which two architecture decisions should you recommend to satisfy these business requirements? (Select TWO.)
- Execute the video transcoding jobs using Cloud Run jobs configured to leverage Spot VM capacity pricing (or Spot placement where available) / preemptible compute execution.Answer
- Configure Cloud Storage Object Lifecycle Management rules on the raw video bucket to transition objects from Standard to Archive storage after 30 days.Answer
- CDeploy a dedicated, multi-zone Google Kubernetes Engine (GKE) cluster with auto-provisioned standard n2-standard-4 node pools running 24/7 to guarantee immediate processing readiness.
- DPurchase 3-year Committed Use Discounts (CUDs) for dedicated Compute Engine instance types sized to match peak upload batch processing volumes.
Answer
To optimize costs for fault-tolerant, unpredictable video processing and storage, the organization should execute containerized batch jobs using Spot compute capacity and configure Cloud Storage Object Lifecycle Management rules to automatically transition older raw files to low-cost archival storage classes after 30 days.
The combination of using Spot-backed serverless batch job execution and automated Cloud Storage Object Lifecycle Management directly satisfies the business goals of reducing compute costs for fault-tolerant jobs and eliminating manual overhead when archiving aging media assets.
Step-by-Step Solution
Key Concept
Cost Optimization for Batch Compute and Storage Lifecycle Management
Estimated Time:1m 30s