awesome-wasm-runtimes VS watt

Firecracker is a fine technology, but serverless companies have started taking advantage Wasm's faster start-up times for use cases of running Wasm on the server (https://www.youtube.com/watch?v=yqgCxhPAao0). The deny by default security policy makes Wasm a great choice to run your code in isolation, particularly for maximizing hardware resources in the multi-tenant environments these serverless companies operate.

In the past few years, we have seen more use cases of Wasm emerge outside of the browser. JavaScript engines are now just a fraction of the total number of runtimes available. Wasmtime, Wasmer, WasmEdge, wazero are popular ones for non-browser use cases like blockchain, serverless, and edge computing (although Cloudflare uses V8's Wasm engine). WAMR is a popular one for cyber physical/IoT devices. There's a nice list here: https://github.com/appcypher/awesome-wasm-runtimes

There's plenty of other ways to interact with Wasm, most of which are secure. (Wasmtime is the one I'm most familiar with, which is why I linked to it.)

Yeah, this is one of many non-browser runtimes, e.g. see https://github.com/appcypher/awesome-wasm-runtimes

Lunatic is more opinionated than most of these or node, though, in that it's trying to emulate a particular concurrent system design pattern borrowed from Erlang/BEAM.

There's an overview of different WASM runtimes with features: https://github.com/appcypher/awesome-wasm-runtimes

Now, the creation Bytecode Alliance, the development of multiple WebAssembly runtimes and the work of the W3C WebAssembly Community Group is why I belive it will get popular, but the capability-based security model is why I want it to get popular.

⚖ I think it's really cool that a plugin author could compile their C++ to .wasm such that a single plugin binary can run on either Linux or Windows (don't need an x86 .dll, x64 .dll, x86 .so, x64 .so...) and in a sandbox (no arbitrary syscalls or Win32 calls, just the interfaces given to it), while still getting near native AOT speed. Though, it's hard to judge which one to choose from now with all the wasm engines that are available (https://github.com/appcypher/awesome-wasm-runtimes), with wasmtime or inNative being two others I've considered for my project. I'll definitely look into this one though, given it supports many of the newer proposals.

Numerous native runtimes for webassembly already exist[0], with the current popular choices apparently being Wasmer[1] and Wasmtime[2].

All one would need to do (AFAIK) is ship a client for all major platforms, as is done with Electron (and web browsers themselves, and everything else.)

[0]https://github.com/appcypher/awesome-wasm-runtimes

The precompiled binary is not a sandboxed WASM binary. Despite the name "watt" it has nothing to do with https://github.com/dtolnay/watt . You can look at the actual code to see for yourself.

* Related: watt is one approach to pre-compile proc-macro crates using WASM.

In addition to thiserror and anyhow, our resident superhuman Rust-improving Robot, dtolnay, also developed an experiment in distributing precompiled proc macros as WebAssembly named Watt and, though I never bothered to create a Zulip account so I don't know what was said, I'm told there has been discussion around the idea of implementing something in that vein.

I think when we have Cranelift, Mold, and maybe Watt all working together then compile times will basically be a non-issue. It'll be a few years though.

Check out https://github.com/dtolnay/watt - it's a really interesting solution to the problem!

I really like the idea of Watt: https://github.com/dtolnay/watt Run macros in a wasm sandbox so they can't touch anything you don't explicitly allow.

I really wish there was more interest in getting something like Watt upstreamed.