5X Faster Edge Functions: V8 Isolates To Firecracker MicroVMs

TL;DR

Prime Big Deal Days · Oct 6–7Offer from Amazon

Get the latest gadgets delivered free — and shop member deals

• Fast, free delivery on millions of items
• Access to Prime Big Deal Days deals on October 6–7
• Prime Video, Amazon Music and more included

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Netlify has moved Edge Functions from an external hosted execution service to Firecracker MicroVMs on its own edge network. The company reports median warm invocation times of about 5–6 milliseconds, compared with 25–40ms before, while saying the change does not require users to alter their functions.

Netlify has moved Edge Functions from an external hosted execution service to Firecracker MicroVMs running inside its own edge network, a change the company says reduces median warm invocation time to about 5–6 milliseconds from 25–40ms. Netlify says existing functions and development workflows remain unchanged, while the new setup is intended to improve performance, isolation and reliability.

The reported latency figures cover a warm invocation: routing a request to a compute node, entering a MicroVM, running the function and producing response headers. Netlify reports a median time of roughly 5–6ms, down from 25–40ms on its previous infrastructure. The company also reports that p99 invocation times are 47.4% faster, availability is 99.998%, and edge-function log delivery is five times faster. These are figures supplied by Netlify; the report does not provide an independent measurement or a detailed measurement period.

Netlify says a function request first reaches its nearest edge node, where the system checks whether the request path matches a deployed function. Matching requests are sent to a compute node in Netlify’s network. A machine specification identifies the runtime, platform and function images, along with resource and connection limits. Netlify says deploys with different code or environment variables receive separate service identities and do not share a MicroVM.

The company says a cold invocation occurs when a region has not yet seen a function and the required images must be fetched. This accounts for about 1.2% of invocations and takes about 9ms on average, according to the report. Netlify says customers can continue using URL imports, npm packages, Node built-ins, netlify.toml declarations and local development as before.

Faster Requests Without Code Changes

Edge functions run before a site responds to matching requests, so their execution time adds to the wait for a user. A reduction in warm invocation latency could benefit sites using functions for personalization, routing and authentication, particularly when those functions run frequently. Netlify says its platform handles about a billion Edge Function executions a day, though the report does not give a measurement period for that figure.

The architecture change also addresses where and how customer code runs. Keeping execution within Netlify’s network removes the previous trip to an external hosted service, according to the company. Netlify says separate MicroVMs for different deploys provide stronger isolation than the V8 isolates used by its previous setup. The report describes this as a security and reliability improvement, but does not include independent security testing or comparative incident data.

From Hosted Execution to MicroVMs

Netlify’s report describes a rebuild conducted over several months with Unikraft, which also published an account of the work. Previously, requests matching an Edge Function left Netlify’s network to run on a hosted execution service, then returned to Netlify to continue toward the site’s response. In the new design, the request is routed to a compute node within the company’s edge network.

Netlify uses rendezvous hashing to direct requests for a service to the same compute node. The company says this routing helps keep the service warm and its images available in local storage or cache. It also describes a trade-off: consistently routing a busy function to one node can create a hotspot. The supplied report ends while discussing that issue and does not include the remainder of its explanation or mitigation details.

““Today, they run on MicroVMs inside our own edge network — roughly 5x faster at the median.””

— Netlify

Performance Claims and Remaining Trade-Offs

The performance, availability and log-delivery numbers are Netlify-reported results. The source does not state the dates or duration of the measurements, the number of requests analyzed, the regions covered, or whether the comparisons were produced under equivalent traffic and workload conditions. It also does not provide third-party verification.

Netlify says cold invocations account for about 1.2% of requests and average about 9ms, but does not break that figure down by region or describe how it varies by function size. The report also identifies the possibility of a compute-node hotspot from sticky routing, without enough detail in the supplied material to establish how frequently that occurs or what controls address it. The company’s claims about improved security and reliability are not accompanied here by comparative incident or security-test data.

Further Details on Scaling and Hotspots

Netlify’s report presents the migration as a completed change to the infrastructure behind Edge Functions, with existing customer workflows continuing unchanged. The next useful information for customers would include longer-term latency and availability data, regional performance breakdowns and more detail about how the platform handles overloaded compute nodes. Netlify has not announced a separate customer migration step or a timetable for additional changes in the supplied material.

Key Questions

What changed in Netlify Edge Functions?

Netlify moved function execution from an external hosted service to Firecracker MicroVMs inside its own edge network, according to the company.

How much faster are the functions?

Netlify reports median warm invocation time of about 5–6ms, compared with 25–40ms on the previous infrastructure. It also reports p99 invocations are 47.4% faster; the source does not specify a measurement period or provide independent verification.

Do developers need to change their code?

Netlify says no. It reports that existing approaches, including URL imports, npm packages, Node built-ins, netlify.toml declarations and local development, continue to work as before.

What happens during a cold invocation?

If a region has not previously run the function, the system fetches the required images before execution. Netlify says this happens on about 1.2% of invocations and takes about 9ms on average.

Are the performance and security claims independently verified?

The supplied material presents these as Netlify’s claims. It does not include independent performance testing, security audits or comparative incident data.

Source: hn