Show HN: Iron, a fast middleware-oriented Rust web server framework | Hacker News Reader

Show HN: Iron, a fast middleware-oriented Rust web server framework

github.com

81 pointsjonreem45 comments

I've been waiting to see some Rust frameworks emerge. Thanks for putting this together!

Would you be willing to contribute a test implementation to our project [1]?

[1] http://www.techempower.com/benchmarks/#section=code

This looks like a great resource!

Unfortunately, I'm not sure if anything in Rust should be considered `production-ready` yet, and I haven't worked with any ORM wrappers in Rust (yet).

I think we could get tests 1 and 6, if you'd be interested. I'll try to get to those soon.

I might help out with this too -- I've been learning Rust recently and I suspect this would help me get up to speed a bit more.

That sounds great, find us anytime on the irc #iron if you want any advice or input.

"Iron averages 2500-3000+ requests per second for hello world"

For a hello world in my laptop I can get typically 30k requests with Go, and with mono/httpListener ~10k.

I expected it to be much faster.

Since the tests are relative they should include a script that tests the different frameworks/languages. It shouldn't be too complicated.

The benchmarks are a bit lower than we wanted. I think this is mostly due to an immature ecosystem for doing basic http - working off of c bindings to a proven, fast framework would probably speed this up by a few times.

It does sound odd that you compare your framework to NodeJS framework. Comparing to Java frameworks (a fast language which can do both parallelism and concurrency) would be much more relevant.

To be honest, it's mostly because that's what we had experience with.

The comparison with Go is really interesting because I don't think its http server does anything particularly fancy (like calling across FFI to non-Go code) and I don't see any obvious bottlenecks in rust-http that would cause an order of magnitude less throughput. Would be interesting to see some profiling.

As far as I can tell (and have been told), rust-http starts a new task (thread, here, really) for every request.

Edit: but see https://news.ycombinator.com/item?id=7978299

Maybe redesign it to start some thread early, and put them in a pool.. then dispatch to them.. i think i've saw a thread pool already backed in rust std or in servo.. dont recall right.. i think servo even have a fancy work stealing thread pool..

Also the design of Nginx is sophisticated, because it uses the event io + a thread dispatch design.. (at least it was like this last time i've hacked it) and worth being copyied

By the way, something is up with the server serving the static doc. The link to the docs is wrong, eg https://github.com/iron/router should link to http://docs.ironframework.io/router/, instead of http://docs.ironframework.io/core/router. To add insult to injury, http://docs.ironframework.io/core/router returns 200 with an empty body instead of 404.

That said, the framework itself looks interesting, though the 'iron/alloy/etc' terminology is a little bit of an annoyance.

Thanks for the heads up. We reworked our folder structure after writing the READMEs, and it looks like we forgot to update. I'll get working on that.

The 200 is being served as the default of `rust-http`. Static will never serve a 404 - it will only defer to the next middleware, and the docs server is overly simple, so it just goes to default.

This should be fixed. Thanks again.

Same here, without even looking into what is slow, Clojure with Compojure does 7500 req/s. If Rust would like to be the next close to bare metal language we need some serious performance improvements.

I am not surprised. Most of the machinery necessary for such things is still developing in Rust.

When I was testing my now-discontinued framework (widmann) using a very early version of rust-http, I felt lucky when it didn't drop or break on half of the requests. Reliability comes before performance in this regard and Rust is a bit from done. I expect that to develop quickly once the general shape of code is known and people start profiling and optimizing.

At that time, Rust’s TCP support was flaky and easy to crash; it was the issue normally, not rust-http.

That wasn't a complaint, just an illustration on how things are quickly improving. Yes, it was a TCP issue, but presented to me through rust-http.

In our benchmarks our raw TCP performance has generally been best-of-breed. I suspect the issue is rust-http, which is not an official package.

I suspect the issue is Iron, though rust-http certainly isn't helping much. Here are some graphs, generated from data made by rust-http's benchmarking script (https://github.com/chris-morgan/rust-http/tree/master/compar...), on my box.

A spreadsheet: https://docs.google.com/spreadsheets/d/1-5_4wPPKNBYZ6yuw-rRl... The raw data: https://gist.github.com/cmr/3648bcc6365c1dba83f1

The charts themselves: http://imgur.com/gw1mp40,Hn3929i,h2OQiTh

Software versions:

rustc 0.11.0-pre (44ec28cfac9fa3f738e0e77ccca1d804125fd1dd 2014-07-01 21:26:33 +0000)

go version go1.3 linux/amd64

node v0.10.29

Apache 2.4.9

wrk 3.1.0

Arch Linux, x86_64, CPU: https://gist.github.com/cmr/7cc6df806caff2253667

It's worth mentioning that I ran this under -c 100 or 200, not 1-8, which may change things significantly.

I just re-ran benchmarks using wrk instead of ab and I'm getting 16k req/s instead of 2k. Very strange.

Did you try changing the runtime to libgreen? It uses a thread pool and dispatches jobs to them instead of spawning a new thread for each job.

The Iron team is here. Feel free to ask us anything!

I've been curious about the HTTP implementation, particularly considering that `rust-http` is in bugfix-only mode and `teepee` is still in design phase. Did you folks build it from scratch?

We built Iron on top of `rust-http`, but we're excited for `teepee` too, and there are a few others popping up as well.

Are you planning on lending assistance to Teepee? We are kind of in limbo in terms of http right now and it would be a shame for Teepee to turn out to be vapourware :(

(A hint for any folks experienced with http - this could be your chance to make lasting mark on the Rust ecosystem!)

Not that experienced, but enthusiastic -- does that count ;-)?

Teepee is getting back on track now; I just had a few weeks where I was unable to do much with it, but I started designing and developing things again yesterday.

Simple question. According to [1], middlewares are cloned for each request. Is there a good built-in way of accessing shared mutable data, such as a connection pool?

1: http://docs.ironframework.io/iron/trait.Middleware.html

I would guess `Arc<Mutex<T>>`, `Arc<RWLock<T>>` and `Arc<SomeConcurrentContainer>` would all work (cloning them is a ref-count - the handle gets copied, not the contents).

You are looking for https://github.com/iron/persistent, which offers two utility middleware for exactly this.

Very nice, thanks.

We have a middleware for that! You can look at http://github.com/iron/persistent, which does wrap an RWLock in an Arc. It also provides a struct to make your middleware Share, so you can do similar things.

We also have session, which works on a HashMap under an Arc, although it isolates the mutable data to a session, instead of sharing it across isolated requests.

Do you plan to support handling XML? I mean XSLT transformation, XPath routing, XQuery?

Please don't!

This looks very nice. I've been following rust and go for a while now and I finally have a small service to write that would be a good fit for either language so I'm going to try both.

I noticed one of the repo's more recent issues was caused by a change upstream. The upstream rust devs seem very accessible, but how often do you encounter these kinds of breaking changes. Acknowledging that rust is pre-1.0, is there any way to stay up to date on these kinds of changes before they hit your app so you aren't blind sided?

We spent a lot of time talking with rust devs on #rust, which is full of extremely helpful, extremely nice people who help you as much as possible.

Usually we woke up in the morning to everything broken due to changes in the nightly, but it's pretty easy to grep for [breaking change] and work from there. Most changes take moments to fix.

We see these changes every other day or so, but they're usually small syntax changes, and the community's pretty quick to update.

A good way to stay up-to-date is always This Week In Rust, or /r/rust.

I know rust is pre-1.0 but wow that's a lot more breaking changes than I expected...

The rate of change slowed, but then, as we get closer to 1.0, the rate of change has sped back up again. We need to break as many things as possible as soon as possible, so that we can give the community some time to try it out before we christen a 1.0 release. Does that make sense?

Seems to me like you need at least a year before beginning to think about releasing 1.0.

Naw. The language itself is getting smaller and smaller over time. Most breaking changes these days (though not all!) are in libraries, which have individual stability markers, and the stability of the standard library doesn't block 1.0.

A year is a very, very, very long time in software. Why do you suggest that long? I'm curious.

Possibly so of you break something in a bad way, you have time to find out, break it again, and then make sure that change is a good one.

Reply on news.ycombinator.com