Estimate CDN Bandwidth (GB per month) from traffic
CDN pricing is often expressed as $ per GB. That means your first job is to estimate GB per month delivered from edge to users. This guide gives three practical methods and a validation checklist so your estimate is usable for budgeting.
0) Decide what you are measuring (edge bandwidth)
For CDN bandwidth, the relevant number is bytes delivered from the CDN edge to end users. Do not confuse this with origin egress (cache fill), which is a different leg.
Related: origin egress vs CDN bandwidth.
Method 1: Use CDN analytics (best)
If you already run a CDN, your provider usually reports bytes delivered per time window. Use a 30-day window if available; otherwise multiply a representative day by ~30.4.
- Look for metrics like "bytes served", "data transferred", or "edge egress".
- Prefer a representative window that includes normal days and a peak day.
- If the UI shows GiB or TiB, convert carefully with Unit converter.
Method 2: From RPS and response size (early-stage planning)
If you know request volume and typical response size, you can estimate bandwidth. This works well for API-like traffic, but you must separate heavy-tail endpoints (downloads, large images) instead of blending everything.
Tools: RPS to monthly requests, Response transfer.
- GB/month ~= requests/month × avg response size (GB)
- Split into at least two buckets: API-like (small payloads, high RPS) and download-like (large payloads, lower RPS).
Method 3: Convert from throughput (Mbps)
If you only have throughput charts (monitoring dashboards, load balancer metrics), convert average Mbps into GB/month. Use an average value over time, not peak throughput.
Tool: Unit converter.
- Rule of thumb: 1 Mbps sustained ~= 324 GB/month (decimal GB).
- Example: 50 Mbps average ~= 50 × 324 ~= 16,200 GB/month (~16.2 TB/month).
What to include (and what to model separately)
- Include: HTML, JS/CSS, images, downloads, video segments, API responses served from CDN.
- Model separately: synthetic monitoring, internal traffic, and large one-off migrations.
- Remember: compression reduces bytes, but not request counts.
Common pitfalls
- Bits vs bytes: Mbps (bits) vs MB/s (bytes) is an 8× difference.
- GB vs GiB: decimal vs binary units; convert consistently.
- Heavy-tail endpoints: one download route can dominate bytes while most endpoints are tiny.
- Peak windows: model baseline + peak; do not average peaks away.
Next: turn GB/month into dollars
After you have GB/month, estimate cost with CDN bandwidth cost calculator, then add request fees with CDN request cost calculator. If you want a single page that bundles both, use CDN cost calculator.
How to validate
- Validate with a real 7–30 day analytics window once the CDN is live.
- Validate top endpoints by bytes (not just by requests).
- Validate unit conversions (Mbps vs MB/s, GB vs GiB) before trusting the number.
- Validate that the estimate matches directionality when traffic changes (deploys, incidents, marketing spikes).