Hacker News Reader

Comment by edude03 | Hacker News Reader

Hmm, I might be missing something here, but don't most high performance network applications skip the kernel for this exact reason? (IE http://highscalability.com/blog/2014/2/13/snabb-switch-skip-...)

Makes me wonder how often bypassing the kernel is used in production networked applications.

View on HN

It's built into FreeBSD: https://www.freebsd.org/cgi/man.cgi?query=netmap&sektion=4

Netmap is commonly used in HFT as well as packet filtering applications. I believe Verisign is running some of the root DNS servers with netmap as well, getting millions of connections per second.

Yeah it is, Robert Graham gave a great talk about it over at https://www.youtube.com/watch?v=D09jdbS6oSI with a writeup at http://c10m.robertgraham.com/2013/02/c10m-in-1990s.html

He's "cheating" by using sendmmsg() to send many messages per system call, reducing the number of context switches.

Barely. On a reasonably configured kernel (you need both syscall auditing and context tracking turned off, which is doable at compile time or runtime), a modern CPU should be able to round-trip a syscall in under 40 ns. That only eats 4% CPU at 1M syscalls per second.

(It's slightly worse than that due to extra cache and TLB pressure, but I doubt that matters in this workload.)

There is a project "Kerlnel" http://kerlnel.org/ that is supposed to be Erlang instance running on bare metal. Site appears to be down and the Github https://github.com/kerlnel hasn't been touched since 2013. Then there is http://erlangonxen.org/ which puts Erlang on top of Xen instead of another operating system.

Well he should have compared this with onload enabled. He was using a solarflare after all. Limited benefits on mc compared with TCP though.

Yes, given the solarflare I was expecting the article to end up with a receiver coded against ef_vi, which exposes the NIC memory directly (but you have to do the IP/UDP yourself)

Yea, solarcapture is essentially, i think, ef_vi consumer off to another thread.

Lots of High Frequency Trading (HFT) applications use an userland based IP stack.

Reply on news.ycombinator.com