[e2e] queue averaging introduces delay

Ramakrishna Gummadi ramki at aciri.org
Sun Aug 5 14:02:17 PDT 2001


Hi
	We have been thinking about the same for sometime now---a PI
controller has a transfer function of the form (K_p+K_i/s), while the
averaging part of a queue has the form: K/(K+s). It seems to me like the
queue averaging component can subsume (is more general than) the integral
part (assuming a suitable gain factor is used when the averaging component
is used).

thanks,
ramki

On Sun, 5 Aug 2001, Alhussein Abouzeid wrote:

> 
> Kris,
> 
> I haven't read your recent papers in too much detail, but I recall that
> you proposed to improve the dynamics of RED in your recent Infocom
> papers by introducing proportional-integral controllers, the integration part
> obviously introducing some sort of averaging, so your comment
> here (using the instantaneous queue rather than the average) seems to be
> conflicting with your previous work. Maybe you can clarify?
> 
> Thanks,
> Hussein.
> 
> On Sat, 4 Aug 2001, C.V. Hollot wrote:
> 
> > i agree with you all, using instantaneous queue is a step in the right
> > direction ... there is no need to introduce additional "averaging" in this
> > feedback loop ... the dynamics of  AIMD window control and buffer provide
> > enough.
> >
> > kris
> >
> >
> > >From: "Atiquzzaman, Mohammed" <atiq at ou.edu>
> > >To: "'Saverio Mascolo'" <mascolo at poliba.it>,
> > >    Ramakrishna Gummadi
> > >          <ramki at aciri.org>, end2end-interest at postel.org
> > >Subject: RE: [e2e] a new paper on Adaptive RED
> > >Date: Fri, 3 Aug 2001 16:01:29 -0500
> > > > In my opinion, main reason because RED does not work is that
> > > > queue average
> > > > introduces delay for which the discard is no more early as it
> > > > should be.
> > >
> > >
> > >We have observed a similar effect. To solve the problem we came up with a
> > >new algorithm which exponentially reduces the average queue length if the
> > >instantaneous queue length has been low for a certain amount of time. The
> > >technical report is available at
> >
> 
> 




More information about the end2end-interest mailing list