Skip to Content.
Sympa Menu

transport - RE: [transport] XCP reference

Subject: Transport protocols and bulk file transfer

List archive

RE: [transport] XCP reference


Chronological Thread 
  • From: "Injong Rhee" <>
  • To: "'stanislav shalunov'" <>
  • Cc: <>
  • Subject: RE: [transport] XCP reference
  • Date: Fri, 10 Dec 2004 11:41:29 -0500



> -----Original Message-----
> From: stanislav shalunov
> [mailto:]
> To:
>
> Subject: Re: [transport] XCP reference
>
> "Injong Rhee"
> <>
> writes:
>
> > Here is the reference about XCP.
> >
> > http://www.isi.edu/isi-xcp/share/extended.pdf
>
> Injong,
>
> Added the reference along with the reference to the internet-draft and
> replaced the old text (``XCP can only be used if all routers on a path
> support it, making it impractical in the near future'') with the
> following:
>
> XCP can only work if the bottleneck router supports it. (The
> question of incremental XCP deployment appears to be open
> \cite{falk-xcp-00,zhang-2004-xcp}, but the bottleneck router
> will need to support XCP if XCP's control is to be effective.)
> XCP will only become useful for the purposes of building a
> deployable tool, then, when sufficient number of potential
> bottleneck routers support it. As the IETF work on the XCP
> specification is still in its early stages, most routers in
> the field are unlikely to support XCP in the next few years.
> Thus XCP, while an excellent congestion control scheme, does
> not appear directly relevant to the need of building a file
> transfer tool that would be usable in the near future.

I would tone it down a little here since it is NOT a proven fact that the
bottleneck has to be an XCP router. But there is empirical evidence for
that. Also we are necessarily building a tool just for file transfer. Based
on these arguments, I would change the text as follows (it is just my input
-- you can change it):

There exists empirical evidence \cite{zhang-2004-xcp} that XCP is not
effective if the bottleneck router does not support it. The issue of
incremental XCP deployment is still open \cite{falk-xcp-00}. As the IETF
work on the XCP specification is still in its early stages and most routers
in the field currently do not support XCP, XCP, while an excellent
congestion control scheme, does not appear to be immediately useful to our
target community.

>
> Does this sound any better?
>
> --
> Stanislav Shalunov http://www.internet2.edu/~shalunov/
>
> This message is designed to be viewed in boustrophedon.




Archive powered by MHonArc 2.6.16.

Top of Page