RE: [AVT] I-D ACTION:draft-ietf-avt-tcrtp-08.txt

["Yaakov Stein" <yaakov_s at rad.com>]

While I completely agree with the motivation behind this draft,
I have problems with the proposed technique.

In section 1.1 we find:

   While some voice-over-packet technologies such as Voice over 
   ATM (VoAAL2, [I.363.2]) and Voice over MPLS provide bandwidth 
   efficiencies because both technologies lack IP, UDP, and RTP 
   headers, neither VoATM nor VoMPLS provide direct access to voice-
   over-packet services available to Voice over IP.  

The savings of VoATM and VoMPLS (and now ITU Y.1414) are mainly 
due to multiplexing multiple channels into a packet with a single
header,
and not merely to not having the RTP/UDP/IP headers.
Indeed, the significance of the overhead diminishes with the number
of multiplexed channels.

So to solve the efficiency problem we need a simple method to multiplex
multiple
channels into a single packet, such as the mechanisms specified in
Y.1414.

The TCRTP draft indeed uses standard mechanisms to multiplexing
channels, 
but, its mechanisms are needlessly complex 
(compression techniques to reduce the overhead, then PPP muxing to
combine them, 
then placing the resulting stream into an L2TP tunnel, and then
performing a further 
header compression on the L2TP/IP headers). 

Why should we go to such lengths when much simpler mechanisms can be
used?

Y(J)S

_______________________________________________
Audio/Video Transport Working Group
avt at ietf.org
https://www1.ietf.org/mailman/listinfo/avt