[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: new WRR sched, updated TBF and "real" ingres implementation



Hello all,
Well very interesting stuff, especially the classful tbf. As for the WRR
can not this queing behaviour being simulated with tbf qdiscs attached
to a prio one? On the other hand I imagine that WRR does not program a
timer for each queue but only when the whole tbf throttles. That way, by
not introducing timer interrupts for every class, as for the prio->tbf's
case, is more computationally efficient.
Panagiotis

devik@cdi.cz wrote:

> Hello,
>
> I needed in my company to distribute bw from shared
> wireless link. Because I was not able to setup CBQ
> for it I patched kernel and implemented new qdisc.
>
> - TBF now supports inner qdisc
> - new WRR (weighter round robin) qdisc
> - hack to allow ingres trafic queueing/shaping
>
> All is based on 2.2.15 kernel, if someone will
> find it useful let me know and I can try to implement
> in into latest kernel.
> All infos are at http://luxik.cdi.cz/~devik/qos.htm .
>
> devik

--
-----------------------------------------------------
Panagiotis G.Stathopoulos
National Technical University of Athens
Telecommunications Laboratory
E-mail: pstath@telecom.ntua.gr
Tel: +30 1 772 1479     Fax: +30 1 772 2534
------------------------------------------------------