3 ms·
Eh. Async and to a lesser extent green threads are the only solutions to slowloris HTTP attacks. I suppose your other option is to use a thread pool in your ser
by josephg 3y ago
Eh. Async and to a lesser extent green threads are the only solutions to slowloris HTTP attacks. I suppose your other option is to use a thread pool in your server - but then you need to but hide your web server behind nginx to keep it safe. (And nginx is safe because it internally uses async IO).
Async is also usually wildly faster for networked services than blocking IO + thread pools. Look at some of the winners of the techempower benchmarks. All of the top results use some form of non blocking IO. (Though a few honourable mentions use go - with presumably a green thread per request):
https://www.techempower.com/benchmarks/ https://www.techempower.com/benchmarks/
I’ve also never seen Python or Ruby get anywhere near the performance of nodejs (or C#) as a web server. A lot of the difference is probably how well tuned v8 and .net are, but I’m sure the async-everywhere nature of javascript makes a huge difference.
- klooney 3y agoAsync's perfect use case is proxies though- get a request, go through a small decision tree, dispatch the I/O to the kernel. You don't want proxies doing complex logic or computation, the stuff that creates bottlenecks in the cooperative multithreading.
- seabrookmx 3y agoMost API's (rest, graphql or otherwise) are effectively a proxy. Like you say, if you don't have complex logic and you're effectively mapping an HTTP request to a query, then your API code is just juggling incoming and outgoing responses and this evented/cooperative approach is very effective.