[dxhdr]

What problems is WASI 0.2 and the Component Model solving? I've read the project goals but I still don't quite understand.

I mean, I guess having some defined interface between languages might be useful for some specific usecases, I just don't see what that has to do with WASM and I also don't understand how this complexity would be a net benefit to the broader ecosystem.

[pjmlp]

Basically since they want to redo P-Code/M-Code/JVM/CLR/BEAM/TIMI/Language Environments/..., now they need to re-invent CORBA, DCOM, XPC, AIDL, gRPC as well.

They already reinvented application servers with WASM containers on Kubernetes.

Otherwise how one would expect to create startups whose goal is to deliver services on top of WebAssembly, something whose original goal was only being a safer ActiveX, Flash, Silverlight, Applets, PNaCL, asm.js, alongside JavaScript.

[flohofwoe]

Exactly. AFAIK the idea is the same as 1990's COM or CORBA, but for WASM (e.g. being able to call between WASM "DLLs" written in different languages but without having to define a C API for each DLL), so pretty much irrelevant for statically linked WASM apps or anything that uses C APIs anyway to talk to each other.

[jeffparsons]

> pretty much irrelevant for statically linked WASM apps

One benefit off the top of my head is the extra isolation between component instances within this statically linked app: you can be shielded from many of the ways libwhatever might ruin everything because its linear memory can be totally isolated from the root component (and vice versa) with surprisingly little overhead.

The Wasm Component model takes security seriously, much in the way that Core Wasm does. Even without everything else it has to offer, this is something our craft is desperately in need of.

[pjmlp]

If Core Wasm took security seriously, the memory access opcodes for arrays would be bound checked like on MSIL and TIMI, instead of "as long as it is inside of a linear memory segment we're good" approach.

[flohofwoe]

Arguably, memory safety inside the WASM heap is the job of the compiler, it's often possible to omit runtime range- and null-checks with the control-flow knowledge the compiler has access to which would be tricky to transmit down to the bytecode level (and even then it would only benefit a handful of languages that have builtin concepts for "ranges" and "nullability".

If you need more fine-grained isolation, split your code into different WASM instances.

[pjmlp]

"Arguably, memory safety inside the UNIX process heap is the job of the compiler, ...."

See what I did there?

And then we are back at the root problem of goverments having to step in, and hardware vendors coming up with hardware memory tagging, because that isn't scaling.

[flohofwoe]

> "Arguably, memory safety inside the UNIX process heap is the job of the compiler, ...."

In my mind there's nothing controversial about this statement at all.

If memory corruption inside the sandbox allows to manipulate anything outside the sandbox, then the sandbox is broken. It's as simple as that, and it doesn't matter whether it's the WASM heap or a process heap.

[nilslice]

and CM copies everything everywhere between components - so this fictional need to build your program out of wasm components made of several languages which in turn also may link many components will just be a nightmare for performance and memory. WIT has to have a “borrow” concept to alleviate the pain.

[leucippe]

There's in-progress design work in SIG Embedded to eliminate unnecessary copies, because that will help using interfaces on embedded devices.

[leucippe]

For background, WASI 0.1 is one interface: a smaller POSIX. Translating POSIX's low-level ABI concepts into WASM is enough for this. WASI came out of WASM's goal of portability, and 0.1 makes WASM portable across POSIX systems.

WASM and WASI aim to support other embeddings. WASI 0.2 introduces an interface for HTTP servers and clients, aimed at cloud/edge. Currently Docker images (on an OS) get used here, but POSIX, even small POSIX, is bloat. WASI 0.2 plans a range of other interfaces for embedding WASM e.g. IOT, so that even without POSIX, you still have portable standards, in your preferred language.

HTTP is a high-level idea. HTTP libraries in any language are never about writing binary to TCP sockets, they're about requests and responses. Depending on sockets would miss the point of leaving POSIX behind. A smaller, loosely-coupled API surface gives both sides more freedom. The runtime can work with the system, the language can work with its users, and the tooling can generate minimal glue that links them together.

Why define WASI APIs with an abstract interface rather than a rigid binary one? Because it follows the spirit of WASM's existing goals: low-level when needed for performance, and high-level whenever it makes WASM simpler to target, embed or reason about. That's why WASM has low-level assembly instructions yet high-level structured control-flow. It's also why WASI uses types appropriate for the API in question (e.g. files in the POSIX interface, responses in HTTP) yet translates them to binary glue when linking real hosts and guests.

Lin Clark's 2019 article[1] raises an issue that comes with low-level interfaces: But if a function takes or returns anything besides numbers, things get complicated. You can either: - Ship one module that has a really hard-to-use API that only speaks in numbers… making life hard for the module’s user. - Add glue code for every single environment you want this module to run in… making life hard for the module’s developer. Clark argues WebAssembly is the right abstraction level to solve this issue in an efficient yet ergonomic way. Frankly, if your WASM module written in Go wants to talk to your host in Java, I don't see why that needs to involve C bindings.

The Component Model falls out of WASI as a bonus. WASI needs to interoperate modules with their embeddings. From there WASI may as well (Lin Clark puts it) "interoperate all the things".

[1]: https://hacks.mozilla.org/2019/08/webassembly-interface-type...