ITU-T TELECOMMUNICATION STANDARDIZATION SECTOR   Document AVC-847

Study Group 15                                   October 1995

Experts Group for Video Coding and Systems
in ATM and other Environments

(Rapporteur's Group on part of Q.2/15)

                     Study Group 15 - CONTRIBUTION

Question: 2/15

SOURCE:   IBM

TITLE:    Comments on H.321 and H.310

Purpose:  Discussion and Proposal




ABSTRACT: Comments on H.321 and H.310









H.321 COMMENTS
---------------
ú ITU SG 11 has defined using the Q.2931 Broadband Higher Layer
  Information element for the exchange of terminal Protocol, forward and
  reverse multiplexing capabilities.  The note on page 9 and the table of
  Q.2931 Information Elements on page 15 need to be updated.

ú In Section 5.6.1 "Convergence Sublayer Functions" where the
  requirement to support the Structure Data Transfer mode by all H.321
  terminals are discussed, it would be worth while to highlight that the
  sending of the Structure Data Transfer Pointer in a single VC carrying
  a single ISDN channel (H0, H11 or H12) is driven by compliance with
  I.580.  This gives the reader a little insight into why the SDT is
  required in all three cases on Page 12..



H.310 COMMENTS
---------------
ú The cost and complexity of supporting both native ATM (MPEG-2,
  H.222.1. etc)  and H.320 (H.221TDM multiplexing, etc) modes of
  conferencing in every H.310 terminal is a concern.  We agree that
  Interworking between multiple terminal types is needed, but not at the
  expense of the end station.  MCUs and conferencing gateways can be used
  to amortize the complexity of interworking over vast numbers of end
  stations.  H.310 should support a native only (MPEG-2, H.222.1, etc.)
  terminal profile.

  This appears to be the approach of other recommendations such as
  H.324.



ú In the September 30th draft of H.310, Section 4 "System
  Description"  it is a misnomer to address both Q.2931 and DSM-CC User-
  to-Network signaling in the same box in the terminal configuration
  picture.  DSM-CC UN should probably be identified as a separate
  protocol stack, much in the same way as we would treat out-of-band
  H.245 or the T-series.