Title: Liaison Statement to ITU-R from the IETF Transport Directorate on TCP over Satellite Collaboration
Received: 13 January 2005

Dear Director Timofeev,

We were very interested to read your letter of 13 October 2004 [1]
informing the IETF of a new Study Group 4 Recommendation on
"Performance enhancements of TCP over satellite networks."  Beyond the
issue of whether the Recommendation may include RFC abstracts (we
believe Scott Bradner has already addressed this), we would like to
offer some technical feedback on IETF technologies and activities
which might be of interest to SG4.

First, we should make clear that we believe the list of IETF RFCs
included in your letter seems very appropriate to improving TCP
performance over satellite.  In the spirit of collaboration, offered
in your letter, there are several suggestions we would like to make.

1. Several of the RFCs listed have been obsoleted by new work.  The
   table below summarizes the obsoleted RFCs:

   Obsolete     New
   ========     ========
   RFC2414      RFC3390
   RFC2481      RFC3168
   RFC2582      RFC3782

2. The IETF had a working group on TCP Performance over Satellite
   (tcpsat) which produced two of the RFCs on SG4's list: RFC2488,
   "Enhancing TCP Over Satellite Channels using Standard Mechanisms",
   and RFC2760, "Ongoing TCP Research Related to Satellites."  This
   work was later generalized to encompass any 'disadvantaged' link in
   another working group, Performance Implications of Link
   Characteristics (pilc) [2].  The output from the PILC working group
   will likely also be of interest to SG4:

    o Advice for Internet Subnetwork Designers (RFC 3819)

    o Performance Enhancing Proxies (PEPs) Intended to Mitigate
      Link-Related Degradations (RFC 3135)

    o End-to-end Performance Implications of Slow Links (RFC 3150)

    o End-to-end Performance Implications of Links with Errors (RFC
      3155)

    o Advice to link designers on link Automatic Repeat reQuest (ARQ)
      (RFC 3366)

    o TCP Performance Implications of Network Path Asymmetry (RFC
      3449)

3. As satellite networks tend to be sensitive to efficient use of
   bandwidth, we recommend reviewing the RFCs and works-in-progress by
   the IETF working group on Robust Header Compression (rohc) (see
   [3]).

4. The IETF continues to evolve TCP in the TCP Maintenance and Minor
   Extensions (tcpm) working group [4].  While the group is concerned
   with general TCP modifications, some modifications are motivated by
   specific environments (e.g., wireless networks).  The TCPM WG would
   benefit from the participation and perspective of the members of
   ITU-R SG4 and we encourage their participation.
                                                                
While the TCPSAT working group is no longer active, the mailing list
continues to be maintained with over 1000 subscribers from both the
IETF and satellite communities.  We would like to offer the TCPSAT
mailing list as a forum for additional collaboration between the ITU-R
and IETF on this topic.  Anyone may subscribe to this list.
Instructions for subscribing may be found at [5].

Best Regards,

    For the IETF Transport Directorate:

    Aaron Falk <falk@isi.edu>
    Chair, past TCPSAT and PILC working groups

    Mark Allman <mallman@icir.org>
    Chair, TCPM working group

    Scott Bradner <sob@harvard.edu>
    IETF liaison to ITU

------------------
[1] http://www.ietf.org/IESG/LIAISON/file58.PDF
[2] http://www.ietf.org/html.charters/OLD/pilc-charter.html
[3] http://www.ietf.org/html.charters/rohc-charter.html 
[4] http://www.ietf.org/html.charters/tcpm-charter.html
[5] http://mailman.isi.edu/mailman/listinfo/tcpsat