Re: applicability of Tcp Throughput model

[Chadi BARAKAT <[e-mail]>]

Hi Cathy, look at my IEEE Network paper "TCP/IP modeling and
validation". Also look on the ITC-17 paper "TCP modeling in the presence
of non-linear window growth". I discussed the issue of correlation
between round-trip time and window size, in networks where you have
small number of connections and lot of bandwidth.

Chadi

cathy_nj a écrit :
> 
> Hi all
>   I can not understand the Tcp throughput model very well with their applicability ,anybody help me?
>   There are two models of Tcp throughput,one is Sally 's simple model which do not take into account rate reductions due to Tcp timeouts.The other is proposed by Jitendra Padhye which is the model of RENO.Both models assume that the round-trip time and the loss rate are independent of the estimated rate.
>   The problem is why this assumption will have a result : the models work well in environments with a high level of statistical multiplexing such as internet,but not in the environments with only few flows share a bottleneck link.Does it means in internet the RTT and loss rate are indeed independent of estimated rate? please give me some explanation.
> 
>   Cathy
> 
> ______________________________________
> 
> ===================================================================
> ÐÂÀËÃâ·Ñµç×ÓÓÊÏä (http://mail.sina.com.cn)
> ÐÂÀË·ÖÀàÐÅÏ¢£º¶þÊÖÊг¡×ßÒ»×ߣ¬¸Ã³öÊÖʱ¾Í³öÊÖ£¡ (http://classad.sina.com.cn/2shou/)

-- 
http://www.inria.fr/planete/personnel/Chadi.Barakat