Workflow Cost Optimization

Proposed State — Decision Spectrums

For each decision, present: the variable, how cost scales, what drives the right answer, and a concrete recommendation for the user’s context.

Encoding and source transport

Packaging and origin (MediaPackage v2)

Ad insertion (MediaTailor)

CDN delivery (CloudFront)

Storage (S3)

Cost Amplifiers — Upstream Decisions That Cascade

These decisions multiply cost across multiple downstream services simultaneously. Flag them early because they have outsized impact.

When presenting cost optimization, always trace the decision through the full chain to show total impact, not just the single-service cost.

Procedure

1. Gather context: environment (prod/dev), audience profile, SLA, content type, expected viewership.
2. Query existing resources for baseline costs and utilization.
3. For each applicable decision, state where on the spectrum you recommend and why.
4. Find current per-unit pricing for the relevant services. If the AWS MCP server’s search_documentation tool is available, use it; otherwise consult the public AWS pricing pages for each service.
5. Compute costs deterministically, not in your head — do NOT perform mental math or approximate in prose. If a Python runtime is available (the AWS MCP server’s run_script tool, or a local interpreter), write and run a script that:
   • Accepts the pricing units from step 4 and the user’s parameters (hours, bitrate, viewer count, segment duration, etc.)
   • Calculates monthly cost per service and total
   • Computes all variants in one script when comparing scenarios, returning a comparison table If no runtime is available, lay out the arithmetic explicitly (per-unit price × quantity for each line item) and sum it step by step so the calculation is auditable.
6. Present results as a table with per-service breakdown. If multiple scenarios, show side-by-side.
7. Wait for user approval before creating resources.