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

["Dan Wing" <dwing at cisco.com>]

> > ATM's multiplexing only helps significantly if you can take 
> > advantage of it, which requires chosing a codec and 
> > packetization interval the fit into the user payload, and 
> > requires multiple streams terminating on the same VPI/VCI.
> 
> ITU Y.1414 uses AAL2 directly over MPLS.
> Similarly draft-ietf-pwe3-tdmoip uses AAL2 directly over UDP/IP.
> 
> The format is :
>   PSN (MPLS or IP) headers
>   first connection : mini-header + payload
>   second connection :  miniheader + payload
>   ...
>   last connection : miniheader + payload
> 
> The only payload size restrictions are maximums due to the 
> size of the AAL2 length field in the miniheader.

You're comparing VoMPLS to VoIP, and I was comparing VoATM to VoIP.
The weakness of VoATM is the minimum 53-byte packet size.  VoMPLS
doesn't have that weakness, but VoMPLS still has smaller headers
than VoIP.

> > We needed a compression technique that saved bandwidth even 
> > in the absence of multiplexing -- such as a branch office 
> > doing VoIP with a single user.
> 
> In that case you only need a robust header compression method.

ROHC, by itself, doesn't work if packets are reordered unless 
draft-pelletier-rohc-over-reordering-00 is used, and ECRTP, by
itself, doesn't work over multiple hops -- ECRTP needs to be
encapsulated (tunneled) inside of something else.  And TCRTP
defines one mechanism for such tunneling.

-d


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