A streaming startup receives raw camera uploads in many codecs and must produce HLS renditions at several bitrates. Volume is bursty, with long idle periods. Which option minimizes spend and operational work?
Show answer and explanation
Answer: D
Transcoder API is a managed service that converts media into many formats and renditions, and job templates describe ladders such as HLS outputs at several bitrates. Jobs read from and write to Cloud Storage, and billing is per processed media, so idle periods cost nothing. No servers or scaling logic need to be built or patched by the team.
Why the other options are wrong
- A. A standing GPU pool bills during every idle hour, which wastes money with bursty demand. It also needs OS patching, FFmpeg upgrades and scaling scripts maintained by the team.
- B. GKE can run transcoding, but it adds cluster administration, node pool sizing and a custom scaler that must be built and tuned. Cron-driven scaling reacts slowly to new uploads.
- C. Cloud Functions has execution time and memory limits that large video files can exceed, and bundling a full ffmpeg toolchain makes the function fragile. It is not designed for long, CPU-heavy media jobs.
Exam tip. For managed media processing at variable volume, prefer a dedicated managed API over self-run encoders on VMs or Kubernetes.