Flowtables -- any tuning hints?
adrian at freebsd.org
Tue Jul 14 02:28:29 UTC 2009
2009/7/14 Kip Macy <kmacy at freebsd.org>:
>> The people who need 10gb/s are isps, universities and telcos; all of whom have a large number of flows. So I'm not sure exactly who is going to benefit from the work.
> There seems to be something unusual about the "large number of
> prefixes" crowd in that any facility that doesn't directly benefit
> them is not worth having. You are not the first to step up and sneer
> with contempt, and yet do nothing to address the architectural flaws
> that hamper forwarding performance for your workload, and you will not
> be the last.
As one of "those people", the reason that we (I) dismiss flow table
stuff as not scalable is because in those specific environments, flow
table stuff has been used, abused and dismissed > 10 years ago as just
The problem is this (and I've spoken to Kip at lengths about this too)
- a company funded Kip's work to improve performance in a particular
area and with a particular set of functions/features. Whether or not
_I_ particularly think it is or isn't a good idea isn't really
relevant if I'm not willing to stand up and do something about it.
If people would like to see general packet forward improvements
regardless of the traffic distribution, they should contribute the
work back to the project. Either talk to those who have expressed an
interest (Kip, Jeff); talk to the FreeBSD foundation directly and
consider donating/sponsoring some larger scale project work to get it
done, or fund a developer directly somehow. Anything else isn't
More information about the freebsd-current