hn.today

5x faster Edge Functions: V8 isolates to Firecracker MicroVMs

netlify.com184 points83 comments
Screenshot of 5x faster Edge Functions: V8 isolates to Firecracker MicroVMs

Netlify replaced V8 isolate-based execution with Firecracker MicroVMs for Edge Functions, cutting median warm-invocation latency from roughly 25-40 ms to about 5-6 ms and improving p99 latency by 47.4%. About a billion Edge Function invocations run daily; the new platform also delivers 5x faster log delivery and 99.998% availability. Cold invocations occur on ~1.2% of requests and average ~9 ms while fetching images. The change preserves developer ergonomics (URL imports, npm, Node built-ins, netlify.toml, local dev) but yields markedly better performance, security, and reliability.

The runtime path: a request lands at the nearest edge node, which writes a machine spec naming three images (runtime, platform, function), computes a service ID from the spec hash and site info, and routes to a compute node via rendezvous hashing for cache stickiness. Compute nodes create per-site services, fetch missing images once, and run each function in its own Firecracker MicroVM (created in <1 ms, starting ~2 ms at p99) with EROFS-packed bundles memory-mapped and snapshot/restore used to scale to zero. Isolation now prevents cross-tenant contamination, and operational controls - relaxed stickiness to avoid hotspots, local DNS, expanded metrics, circuit breakers, a control plane, and separate compute/edge fleets - support fast rollouts and resilience.

Read on netlify.com83 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in Web

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.