Re: [rohc] Sigcomp presence-specific dictionary

[cco <[e-mail]>]

On Mon, Mar 26, 2007 at 02:46:07PM +0300, Miguel Garcia wrote:
> Hi Cristian:
> 
> Thanks for your comments. Inline answers.
> 
> cco wrote:
> >On Tue, Mar 06, 2007 at 09:23:50AM +0200, Miguel Garcia wrote:
> >>Hi.
> >>
> >>I have been working on a presence-specific static Sigcomp dictionary 
> >>for SIP presence. The draft is addressed to the SIMPLE WG, which is 
> >>working on SIP presence. But I want to make the ROHC WG aware of the 
> >>draft and solicit comments.
> >>
> >>Until the draft is officially available, you can download a copy from:
> >>
> >>http://people.nokia.net/~miguel/drafts/pre/draft-garcia-simple-presence-dictionary-02.txt
> >>http://people.nokia.net/~miguel/drafts/pre/draft-garcia-simple-presence-dictionary-02.html
> >
> >cristian: hi! the section "1.  Introduction" says at some point:
> >
> >"Sigcomp endpoints will announce the availability of
> >   one or both dictionaries during the Sigcomp initialization phase."
> >
> >what exactly do you mean by "Sigcomp initialization phase"? sigcomp is
> >piggy-backed on SIP in this case and there is no explicit sigcomp
> >"signalling".
> 
> 
> "SigComp initialization phase" refers to when the
> endpoint has not received an acknowledgment from the other end
> that the other end has received the message containing the
> announcement about presence dictionary. Perhaps the text would be better
> phrased as:
> 
>    Sigcomp endpoints will initially announce the availability of one
>    or both dictionaries until the other end acknowledges that it has
>    received the announcement.
> 
> 
> 
> 
> >
> >on the other hand rfc3485, "The Session Initiation Protocol (SIP) and 
> >Session Description Protocol (SDP) Static Dictionary for Signaling 
> >Compression (SigComp)", defining a similar dictionary (sigcomp "local" 
> >state)
> >puts a requirement on the implementation of sigcomp stacks for sip:
> >
> >"The static dictionary is unique and MUST be available in all SigComp
> >   implementations for SIP/SDP."
> >
> >(and the need for such a requirement is also explained in the rfc.)
> >why not having the same kind of req. in your draft?
> 
> 
> We were able to mandate the SIP dictionary for SigComp because they were 
> created at the same time. However, nowadays, there are millions of 
> deployed SIP endpoints that implement SigComp (and the SIP static 
> dictionary), presence, and other services, that do not obviously 
> implement the presence-specific dictionary that we are discussing here. 
> So, obviously, we cannot assume that an implementation will have a 
> built-in presence-specific dictionary, unless it is advertised.

cristian: o.k. I understand; however this will reduce the "initial" 
compression efficiency (compared to a scenario where the dictionary will 
be used right on from the start). 

bye now!
cristian

_______________________________________________
Rohc mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/rohc