RE: Drafting mux WG charter

From: Yaron Goland (
Date: Mon, Feb 15 1999

Message-ID: <3FF8121C9B6DD111812100805F31FC0D08792F11@RED-MSG-59>
From: Yaron Goland <>
To: "''" <>, Joe Touch <>, Ted Hardie <>, Mark Day <>, Rohit Khare <>
Cc:, Josh Cohen <>, Lance Olson <>
Date: Mon, 15 Feb 1999 15:30:07 -0800
Subject: RE: Drafting mux WG charter

My personal interest in MUXing is to re-use a single TCP connection to some
MUX port as a mechanism to simulate arbitrary TCP connections to arbitrary
ports. Put another way (brazenly stolen from Josh Cohen) I should be able to
replace my TCP/IP stack with a MUX enabled stack and none of my applications
should notice any difference. 

In the best of all worlds I would like this issue to be declared explicitly
in-scope even if the core MUX protocol doesn't support it. For example, I
could imagine an extension spec built on top of MUX which would extend the
MUX open command to support specifying a TCP port that the particular mux
channel that was opened should be treated as speaking over.

I believe I understand where you are going with the security language and I
even think I agree. But the current text definitely needs some help. However
enough people are already beating on you about that one that I'm sure you
have gotten the point. =)

> -----Original Message-----
> From: []
> Sent: Monday, February 15, 1999 11:38 AM
> To: Joe Touch; Ted Hardie; Yaron Goland; Mark Day; Rohit Khare
> Cc:
> Subject: Drafting mux WG charter
> I'm working on drafting a charter to work on a multiplexing 
> protocol; such a thing was discussed at the HTTP-NG BOF at 
> IETF-43, at which you spoke up.  If you
> have any interest, I'd appreciate your comments.  Go to
> <> for pointers to the charter and
> instructions for the mailing list where we're discussing it
> (  Some discussion has also spilled over to
>, but please let's henceforth keep it on one list
> only.
> Thanks,
> Mike