Yeah Renata, Newreno modifies Reno in the Fast Recovery phase to perform
better in the context of multiple packet losses in the same window. But my
question was different : why does the FullTcp variant of Newreno
(TCP/FullTcp/Newreno) perform differently than the normal Newreno
(TCP/Newreno) ?

-Sid

On 6/16/06, Renata Vidal <[EMAIL PROTECTED]> wrote:
>
> I didn't read the RFC yeat, but is there some modification in newreno? (
> http://www.ietf.org/rfc/rfc3782.txt) Is there any comments in .cc header
> files?
> There are a lot of working groups about congestion and avoidance... there
> is some difference about it in one of them (look for DCCP).
>
> Regards, Renata
>
>
> On 6/16/06, Siddharth Ramesh <[EMAIL PROTECTED]> wrote:
> >
> >
> > Hi all,
> > I was using FullTcp/Newreno agent and found that it behaved differently
> > from
> > the normal TCP/Newreno in terms of congestion control. I am really not
> > sure
> > about what is causing the difference. This difference is especially
> > obvious
> > when the number of flows sharing a bottleneck link becomes high - the
> > FullTcp flows were performing much better in terms of the throughputs
> > achieved.
> >
> > Anybody has any clue as to why this is happening?
> > Thanks,
> > Sid
> >
>
>
>
> --
> Renata Vidal
> "Triste é o destino de quem tenta vencer as batalhas e ter sucesso nos
> ataques sem cultivar o espírito da iniciativa (A Arte da Guerra – SUN TSU –
> pg 102)"
>

Reply via email to