[rohc] Rohc Open source implementations

Hi 

I'm looking for open source implementations of rtp header compression
schemes e.g. rohc, crtp etc. Rohc at http://rohc.sourceforge.net/
<http://rohc.sourceforge.net/>  doesn't seem to implement profile 0x01
(RTP). I've grepped the rohc mail archive from 2003+ and found open
sigcom, but no other rohc implementations. Are you aware of any other
implementation that I could look at?

Background: 

I'm looking to implement a low overhead rtp multiplex scheme over an
IP/UDP transport but would like to develop an understanding of what is
likely to work well. Rohc is interesting and I'm looking at rfc 4224
which looks at mitigating the effects of re-ordering. However,
implementing the whole specification from scratch is likely to be time
consuming, so I'm looking to focus on a minimal profile 0x1 subset e.g.
just implementing r-mode with multiplexing extensions.  Anybody working
on this? 

Assuming that: 
- there are no other implementations out there 
- I'd like to implement something small that deals with reordering e.g.
only O-mode or only R-mode. 
- taking on board rfc4224 and
draft-kapoor-rohc-profiles-reordering-01.txt 

Then I've drawn the preliminary conclusion that implementing O-mode with
rfc4224 (ext-0 extensions) or
draft-kapoor-rohc-profiles-reordering-01.txt fixes might be a good
approach. Does anybody think R-mode only would be better or sufficient
for a UDP channel?

Regards, 
Simon 

Private reply from Ghyslain Pelletier (LU/EAB)
[ghyslain.pelletier@ericsson.com] 

It is not the first time that we get such enquiries. Unfortunately, I am
not aware of any other rohc implementation than the one at sourceforge,
which as you correctly point out is not complete (and I am not sure it
can be trusted to interop either). Wrt crtp and/or ecrtp, I have never
looked for free implementations so I cannot give you a proper answer
here.

A quick reflection - why do you plan on implementing R-mode? R-mode is
rather sensitive to reordering. RoHCv2 might be simpler and faster for
you to implement a subset:

http://tools.ietf.org/wg/rohc/draft-ietf-rohc-rfc3095bis-rohcv2-profiles
/draft-ietf-rohc-rfc3095bis-rohcv2-profiles-00.txt
<http://tools.ietf.org/wg/rohc/draft-ietf-rohc-rfc3095bis-rohcv2-profile
s/draft-ietf-rohc-rfc3095bis-rohcv2-profiles-00.txt> 

Obviously, it is not ready yet but I see it as a complete solution that
includes only the strict necessary to efficiently compress robustly in
the presence of reordering.