RE: Simple Multicast - building a case for a BOF or WG

[Spencer Dawkins <sdawkins@nortelnetworks.com>]

I only mention this because I was amazed by it - in the BOF on
transport-friendly ESP, several people said they were uncomfortable with
network entities that change the loss rates and interpacket timings of TCP -
they understood this as "changing the semantics of TCP". I Would Have
Thought that "generating ACKs at intermediate nodes before they reflect ACKs
sent by far-end nodes" was more the kind of thing that would be considered
"changing the semantics of TCP", so I was pretty surprised by this (and even
more thrilled that we didn't spend the entire PILC session earlier in the
meeting talking about this).

This isn't a TCP comment - it's a heads-up that Ken is correct when he says
there are people running around loose with very strict ideas about what you
can do before it's a revolutionary change!

Spencer

> -----Original Message-----
> From:	Ken Carlberg [SMTP:carlberg@time.saic.com]
> Sent:	Wednesday, May 05, 1999 2:40 PM
> To:	Jon Crowcroft
> Cc:	sm-interest@cs.ucl.ac.uk
> Subject:	Re: Simple Multicast - building a case for a BOF or WG
> 
> jon,
> 
> > again let me remind you: sm is evolutionary, not revolutionary, so
> 
> while i understand the point of view of the SM authors on this aspect,
> what is considered evolutionary is still a subjective.  whispers from
> others in the ietf hallways tend to feel that simply changing the API 
> is grounds for a revolutionary label.