The Exacting Details of "Hello, World!" in Wasmtime

Nick Fitzgerald - F5

What steps must a WebAssembly engine take to run “Hello, World!”? How is sandboxing implemented? And what can we, as developers targeting and building platforms for WebAssembly, learn from it?

This talk will trace the exacting details of everything — and I mean everything — that Wasmtime does to run a simple WebAssembly program: from compiling Wasm bytecode into native machine code, to the design of runtime “vmctx” data structures, to the implementation of WASI syscalls, and all that sits in between. Learn the arcane knowledge of WebAssembly implementation and you will know the cost of dynamic dispatch in your WebAssembly program, how to maximize your function-as-a-service platform’s WebAssembly instance capacity, and even how to contribute new features to Wasmtime itself.