[e2e] How popular is TCP delayed acknowledgement used ?

Jitendra Padhye padhye at icir.org
Thu Apr 10 10:34:10 PDT 2003


See http://www.aciri.org/tbit/

and specifically

http://www.icir.org/tbit/tbit.ps

for some data regarding use of delayed acks etc. 

- Jitu


> Nagle algorihtm is set as default in BSD-based OS. Nevertheless, it can
> be disable by setsocketopt ( .. TCP_NODELAY...)
> 
> Maybe the paper "Rethinking the TCP Nagle Algorithm," Jeffrey C. Mogul,
> Greg Minshall, which showed at
> http://www.acm.org/sigcomm/ccr/archive/2001/jan01/ccr-200101-mogul.pdf
> is helpful.
> 
> Fu Cheng Peng, Franklin
> 
> 
> -----Original Message-----
> From: alok [mailto:alok.dube at apara.com] 
> Sent: Thursday, April 10, 2003 5:31 PM
> To: end2end-interest at postel.org
> Subject: Re: [e2e] How popular is TCP delayed acknowledgement used ?
> 
> 
> Hi,
> 
> To add to this,
> 
> I would also like to know in each of the possible options, what does one
> define as a "window"?
> 
> -rgds
> Alok
> ----- Original Message -----
> From: Yang Luqing <engp1781 at nus.edu.sg>
> To: <end2end-interest at postel.org>
> Sent: Thursday, April 10, 2003 1:58 PM
> Subject: [e2e] How popular is TCP delayed acknowledgement used ?
> 
> 
> > Hi, all,
> >
> >        I want to know whether in most of the hosts on the Internet, 
> > their
> TCP use the delayed acknowledgement policy or acknowledge every packet
> it receives.
> >       Thanks for your help in advance.
> >
> >
> > Regards.
> >
> > Luqing
> >
> >
> 
> 
> 
> 




More information about the end2end-interest mailing list