Does newreno work as designed ?
Jeffrey Hsu
hsu at FreeBSD.org
Fri Jul 4 06:49:58 PDT 2003
> I'm sitting on a ISDN line right now, and I thought the newreno issues
> had been solved, but by disabling newreno I get a distinctly better
> web-surfing experience.
NewReno is a sender-side only algorithm and web-surfing is mostly a
receiver-side experience. If you gather some quantitative numbers
on bulk transfers or a packet trace, I'd be happy to look at them for you.
> Is newreno working as designed right now
Yes, it should work much better than before. There is one outstanding
bug that I now of, which was reported by Lu GuoHan. A patch for it is
in circulation. You can find it at
http://people.freebsd.org/~hsu/tmp/explicitnewreno.diff
and see if it affects your web-surfing experience.
> and if not, who is fixing it ?
Unfortunately, it appears only a few people care about work on
TCP congestion control, even though it affects a lot more users
and is several factors more important for performance than, say, SMP.
> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
> phk at FreeBSD.ORG | TCP/IP since RFC 956
More information about the freebsd-current
mailing list