Elli: Performance oriented web server written in Erlang
1–10 of 22 posts
Re: Elli: Performance oriented web server written in Erlang
#2I wonder how it compares with all the other high performance Erlang web servers. He seems familiar with them, so I'm curious what requirements or functionality changed that they needed to write their own instead of hacking one of the others.
Re: Elli: Performance oriented web server written in Erlang
#3I wonder how it compares with all the other high performance Erlang web servers. He seems familiar with them, so I'm curious what requirements or functionality changed that they needed to write their own instead of hacking one of the others.
According to my simple 'hello world' AB test, elli is about 40% faster than cowboy.
Re: Elli: Performance oriented web server written in Erlang
#4I wonder how it compares with all the other high performance Erlang web servers. He seems familiar with them, so I'm curious what requirements or functionality changed that they needed to write their own instead of hacking one of the others.
According to my simple 'hello world' AB test, elli is about 40% faster than cowboy.
Why is it so much faster? Is it really that much better code, or are there some tradeoffs in the other ones that Elli approaches in a different way?
Re: Elli: Performance oriented web server written in Erlang
#5Color me disinterested.
Show me a real benchmark. I don't really care if its written in Erlang. I can likely blow that 'Hello World' away by writing directly to a socket.
Re: Elli: Performance oriented web server written in Erlang
#6Earlier quoted context omitted.
According to my simple 'hello world' AB test, elli is about 40% faster than cowboy.
Why is it so much faster? Is it really that much better code, or are there some tradeoffs in the other ones that Elli approaches in a different way?
Less features... still isn't doing dates (which is a known slowdown in Erlang)... connection process because handler process...
Still neat.. what impressed me was some of the linked "helper" projects for doing stuff likes stats.
Re: Elli: Performance oriented web server written in Erlang
#7Earlier quoted context omitted.
According to my simple 'hello world' AB test, elli is about 40% faster than cowboy.
Why is it so much faster? Is it really that much better code, or are there some tradeoffs in the other ones that Elli approaches in a different way?
He wrote a custom webserver based on his very specific needs. It's not a general purpose solution, so not too surprising it's faster for his particular use cases.
Re: Elli: Performance oriented web server written in Erlang
#8Earlier quoted context omitted.
Why is it so much faster? Is it really that much better code, or are there some tradeoffs in the other ones that Elli approaches in a different way?
He wrote a custom webserver based on his very specific needs. It's not a general purpose solution, so not too surprising it's faster for his particular use cases.
It'd be interesting to hear about that in more detail - what tradeoffs he made to get that speedup.
Re: Elli: Performance oriented web server written in Erlang
#9what's the difference in architecture to the other "big" erlang webservers (yaws, misultin and cowboy)?
Re: Elli: Performance oriented web server written in Erlang
#10Color me disinterested. Show me a real benchmark. I don't really care if its written in Erlang. I can likely blow that 'Hello World' away by writing directly to a socket.
Hi,
Author of elli here.
It is very hard to create real and meaningful benchmarks. The "Hello world" benchmarks are very useful when writing a webserver and you are curious about where to optimize it, if necessary. Even though they are very superficial, they can also help you compare two webservers.
It is very easy to run the benchmarks on your own hardware. Get elli, then run "elli:start_link()" and hit "/hello?name=john" with apachebench or your tool of choice.
Elli is only useful if you want to write an Erlang application that exposes a HTTP API. If you want raw performance, Haskell has some servers which does 300k+ rps.
Here is a very good article from Steve Vinoski who is overall a very smart and experienced guy on the topic of benchmarking Erlang webservers: http://steve.vinoski.net/blog/2011/05/09/erlang-web-server-b...
Knut