[iccrg] 答复: 答复: [tcpm] New Version Notification for draft-you-tcpm-configuring-tcp-initial-window-00.txt

Hi all, 

A new version has been uploaded. We fix typos and discuss the issue when APP tries to cheat the system by splitting a large flow into small ones.
Your comments are appreciated. 

A new version of I-D, draft-you-tcpm-configuring-tcp-initial-window-03.txt
has been successfully submitted by Jianjie You and posted to the IETF repository.

Name:		draft-you-tcpm-configuring-tcp-initial-window
Revision:	03
Title:		Configuring TCP's Initial Window
Document date:	2014-11-10
Group:		Individual Submission
Pages:		8
URL:            http://www.ietf.org/internet-drafts/draft-you-tcpm-configuring-tcp-initial-window-03.txt
Status:         https://datatracker.ietf.org/doc/draft-you-tcpm-configuring-tcp-initial-window/
Htmlized:       http://tools.ietf.org/html/draft-you-tcpm-configuring-tcp-initial-window-03
Diff:           http://www.ietf.org/rfcdiff?url2=draft-you-tcpm-configuring-tcp-initial-window-03

Thanks,
Jianjie

> -----邮件原件-----
> 发件人: iccrg [mailto:iccrg-bounces@irtf.org] 代表 Runa Barik
> 发送时间: 2014年11月6日 19:56
> 收件人: iccrg@irtf.org
> 主题: Re: [iccrg] 答复: [tcpm] New Version Notification for
> draft-you-tcpm-configuring-tcp-initial-window-00.txt
> 
> Hi,
> 
>    There is a typo in the draft
> draft-you-tcpm-configuring-tcp-initial-window-02.txt.
> 
>                 if s <= theta then
>                    InitCwnd = MaxIW
>                  else
>                    InitCwnd = theta/s * MaxIW + (1-theta/s) * MinIW
>                 end if
> 
>   There is also another draft draft-sallantin-iccrg-initial-spreading-01.txt. I
> think combining both IW function, and  initial spreading may give better
> performance.
> 
> Regards,
> - Runa
> ________________________________________
> From: Joe Touch <touch@isi.edu>
> Sent: Monday, October 27, 2014 5:33 AM
> To: Youjianjie; Runa Barik
> Cc: iccrg@irtf.org
> Subject: Re: 答复: [tcpm] New Version Notification for
> draft-you-tcpm-configuring-tcp-initial-window-00.txt
> 
> On 10/26/2014 8:03 PM, Youjianjie wrote:
> ...
> >> There are no trusted apps in the Internet; there’s no one party to “trust”
> >> them.
> >>
> >> I.e., I have no reason to trust your apps, nor you to trust mine.
> >>
> >> So, ultimately, everyone needs to use the same rules and assume that
> >> everyone will try to game the system.
> >
> > The worst is all the applications use the default maximum IW set by the
> system.
> 
> Please explain the alternative.
> 
> You can't know what the application will do unless you ask it OR you wait until
> the application is done. In the first case, you have to trust the application, but
> applications will lie for better performance. In the second case, you have to
> wait until the application has buffered all its data AND closed the socket to
> know - for sure - how big the window should have been.
> 
> There's really no useful alternative, AFIACT.
> 
> Joe
> _______________________________________________
> iccrg mailing list
> iccrg@irtf.org
> https://www.irtf.org/mailman/listinfo/iccrg