RE: [rohc] Liaison towards 3GPP

["Casati, Alessio (Alessio)" <acasati@lucent.com>]

> Since bandwidth
> is scarse in cellular system, standards for robust header compression
> algorithms are essential to achieve spectrum efficiency. 
> 
I've seen some proposals for header stripping, that is
deliver the RTP payload to the other end of the
wireless channel where the headers are locally reconstructed.

It is true that the network side would have to terminate RTP and
play out real time streams, thus potentially increasing complexity
in the RNC.

On the other hand, one could argue that however 
efficient the Comression 
algorithm could be, stripping the header would 
save some octets on the average, and some $ billions 
- as per Carsten's estimate.

I think we need a ROHC algorithm anyway to run efficiently 
a host of apps, but I'm not sure
that the header stripping proposals won't be the 
preferred way to deliver voice in wireless nets.

Any opinions?


alessio