November '93 IETF: TCP Multiplexing BOF

Jim Barnes <barnes@xylogics.com> Thu, 14 October 1993 19:47 UTC

Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa12228; 14 Oct 93 15:47 EDT
Received: from ietf.cnri.reston.va.us by CNRI.Reston.VA.US id aa28601; 14 Oct 93 15:47 EDT
Received: from ietf.cnri.reston.va.us by IETF.CNRI.Reston.VA.US id aa12215; 14 Oct 93 15:47 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa12086; 14 Oct 93 15:43 EDT
Received: from ietf.cnri.reston.va.us by CNRI.Reston.VA.US id aa28237; 14 Oct 93 15:43 EDT
Received: from [127.0.0.1] by IETF.CNRI.Reston.VA.US id aa12078; 14 Oct 93 15:43 EDT
To: IETF-Announce:;
Sender: ietf-announce-request@IETF.CNRI.Reston.VA.US
From: Jim Barnes <barnes@xylogics.com>
Subject: November '93 IETF: TCP Multiplexing BOF
Date: Thu, 14 Oct 1993 15:43:10 -0400
X-Orig-Sender: mwalnut@CNRI.Reston.VA.US
Message-ID: <9310141543.aa12078@IETF.CNRI.Reston.VA.US>

 Group: TCP Multiplexing BOF (tmux)
 IETF Area: Transport
 Date: Wednesday, November 3, 1993
 Time: 0930-1200
 
 
 One of the problems with terminal servers attached to a LAN is the
 large number of small packets they can generate.  Frequently, most of 
 these packets are destined for only one or two hosts.  The large number of
 small packets causes performance problems not only with the terminal
 servers generating the packets, but also with the hosts receiving the
 packets.  The receiving hosts must pass all these small packets through
 their TCP/IP protocol stacks and send the data to the destination
 applications.  A standard solution to this overhead problem would be 
 useful to both terminal servers and host nodes.  
  
 At the Amsterdam IETF, the CMP protocol (multiplexing on top of the 
 TCP layer) was discussed.  This protocol is described in the Internet
 draft document draft-cameron-cmp.01.txt.  An alternative proposal (known
 as TMux) was presented that multiplexed small packets between the IP and 
 TCP layers.  This protocol is described in the Internet draft document 
 draft-cameron-tmux.00.txt.  The general consensus of those attending
 the Amsterdam BOF was that the TMux proposal was superior to the CMP
 proposal.
 
 The purpose of this BOF is two-fold.  First, to report on implementation
 experience for TMux.  This report will include observed performance of
 the TMux implementation.  Second, to consider whether the TMux protocol 
 definition should be moved forward to Proposed Standard status.
 
 An email list has been set up for discussions of this protocol.  To
 join the list, send a request to cmp-id-request@xylint.co.uk
 
 Please read the TMux protocol document (draft-cameron-tmux-00.txt) before 
 the meeting.  You may also wish to read the CMP protocol document
 (draft-cameron-cmp-01.txt) for historical purposes.
 
 Agenda for the BOF:
 
 - Introduction 
 
 - CMP and TMux:  How we got here
 
 - TMux implementation experience

 - What next?  Should TMux be submitted to the IESG for elevation to
 	Proposed Standard status?