Re: [Tsvwg] draft-floyd-tsvwg-cc-alt-00.txt

[Sally Floyd <[e-mail]>]

Doug -

> The "Difficult Environments" clause is important but pretty 
> non-specific at the moment.   It might be helpful to provide a few 
> more examples illustrating the types of situations you have in mind.  
> It might also worth commenting in some way that the performance of 
> standard TCP is not necessarily all that well understood in some 
> challenging environments -- for example,TCP performance in 802.11 
> wireless environments  is still an active research topic; similarly 
> behaviour with reverse path queueing.  As a result, it might not be 
> entirely reasonable to place a much higher bar on new algorithms than 
> currently exists for standard TCP.
>
> Similar comments apply to the "Responses to Sudden or Transient 
> Events" text.  At the moment its not clear where the real thrust of 
> this is aimed at and so its usefulness as a guideline is perhaps 
> reduced.

We will add more examples to both sections.  We will also try to make 
the context more clear.  The intent is not to place a higher bar on new 
algorithms than on standard TCP,
in terms of performance.   But if a new mechanism is shown to perform 
significantly worse than standard TCP in "Difficult Environments" on in 
"Responses
to Sudden or Transient Events", my hope would be that the IETF takes 
this into account in its recommentations concerning the appropriate 
deployment of the
new mechanism in question.

- Sally
http://www.icir.org/floyd/