Cloudflare Workers vs Vercel Edge TTFB
Status: protocol, not measured result
This methodology compares Cloudflare Workers and Vercel Edge Functions on the same Astro 5 build.
Related pages: benchmark protocol and methodology index.
Test Corpus
- 1 static blog post route.
- 1 ISR-like product page route.
- 1 SSR personalised dashboard route.
- Shared WordPress 6.7+ origin with WooCommerce 9+.
- Identical cache headers and route logic.
Measurement Matrix
| Field | Required value |
|---|---|
| Device profile | Moto G4 |
| Network profile | 4G throttle |
| Run cities | Frankfurt, Warsaw, Amsterdam, Madrid, Stockholm |
| Cache states | cold, warm |
| Runs per route/platform/city/cache state | minimum 5 |
| Primary metric | TTFB |
| Secondary metrics | LCP, cold-start latency, cost per 100k requests |
Result Table
| Platform | Cache | City | Median TTFB | P75 TTFB | Median LCP | Cold Start | Cost / 100k | Notes |
|---|---|---|---|---|---|---|---|---|
| Cloudflare Workers | pending | pending | pending | pending | pending | pending | pending | pending |
| Vercel Edge | pending | pending | pending | pending | pending | pending | pending | pending |
Publication Gate
Do not call this a result until the same Astro build is deployed to both platforms and adapter-specific changes are documented.