Hyperlight: Achieving 0.0009-second micro-VM execution time
This post walks through the demo application from Rita Zhang’s KubeCon NA 2024 keynote, showing how Hyperlight achieves 0.0009-second average micro-VM execution time. It explores the demo app’s architecture — a Rust-based web server with cold and warm VM routes — and explains how Hyperlight creates a brand-new hardware-protected micro-VM for each request, executes a guest function, and tears it down, all within sub-millisecond timeframes. Written by Danilo Chiarlone.