A media streaming company runs short-lived, stateless batch video rendering tasks that can tolerate interruptions and resume cleanly. The resource demands for these jobs fluctuate significantly throughout the day. To minimize compute infrastructure costs for these flexible workloads without committing to long-term resource contracts, which cost optimization strategy should the Cloud Architect recommend?
- Utilize Spot VMs for the rendering compute nodes to take advantage of deeply discounted spare GCP capacity.Answer
- BPurchase 3-year Committed Use Discounts (CUDs) sized to cover the peak capacity required during maximum rendering bursts.
- CDeploy a multi-zone GKE cluster and migrate rendering queue state to Cloud Spanner.
- DGrant the primitive Owner IAM role to the FinOps automated cleanup service account.
Answer
Utilize Spot VMs for the rendering compute nodes to take advantage of deeply discounted spare GCP capacity.
Spot VMs are spare Compute Engine capacity offered at substantial discounts. Because the rendering jobs are stateless and can tolerate interruptions, using Spot VMs maximizes cost savings without imposing long-term financial commitments.
Step-by-Step Solution
Key Concept
Selecting cost-effective GCP compute models for fault-tolerant batch processing workloads.