These benchmarks on
@fastly's Compute@Edge vs Cloudflare workers are seriously flawed.The biggest problem is that they conflate network round trip time with “execution at edge” time. CF is measuring cherry-picked RTT instead of edge compute performance. From their blog: