Re: [tcpm] New ID available: RFC2988bis (RTO calculation)

[Mark Allman <[e-mail]>]

[This is a re-send.  My original note did not make it to the list.  I
 made a tweak to my mail setup.  Lets see if this works. --allman]

> what a quick response :-)

It happens every once in a while.  Don't get used to it! :)

> >  - First, I sent a technical note to the list on Feb/11 about changing
> >    the min.  
> 
> Sure? On this list...?
> Anyway, since you appended the text, I can read it.

In the 'TCP tuning' thread.  I didn't bother to look to see if it
actually made it through.  I found it in my outbox.

> >  - Second, it seems that we should keep these two questions independent
> >    to me.  It seems to me that changing the initRTO is pretty minor and
> >    well-scoped, whereas changing the minRTO has more unknowns.  So, I
> >    am not at all opposed to thinking about both changes as you suggest,
> >    but coupling them is likely to hinder progress on the initRTO
> >    (IMO). 
> 
> Yes, full ACK. However, I suggest that we add some text (like
> your appended note below or your comments here 
> http://www.ietf.org/mail-archive/web/tcpm/current/msg02937.html)
> to the draft, why we not lower minRTO at this time.

It is a fair point that we should perhaps better motivate the 1sec
choice in the document.  A usual problem in IETF documents: we are
better at the 'what' than the 'why'.

allman