Re: [Sip] Question regarding instance ID in GRUU and outbound

[Aki Niemi <[e-mail]>]

ke, 2006-08-09 kello 15:58 -0700, ext Cullen Jennings kirjoitti:
> This email is sort of irrelevant but I'm sending it anyways :-)
> 
> A few year back, I wrote
> http://www.softarmor.com/wgdb/docs/draft-jennings-sipping-instance- 
> id-00.txt
> 
> and it had some use cases for why one might need a long term stable  
> identifier for a device. I think theses still hold. We need something  
> and it needs to survive reboots of a device. You are right it does  
> not matter who generates the ID but the UA needs to store it. Cookies  
> were discussed but rejected due to worries about if there was AT&T  
> IPR on them.

This is pretty interesting. Has anyone verified there is an actual
patent or patent application on the cookie approach?

> I originally proposed something "simple" to generate - I'm a big fan  
> of large random numbers. Others were less comfortable with large  
> random numbers and wanted something algorithmically guaranteed to be  
> unique by some sort of delegated assignment - like MAC or OIDs. Most  
> devices had MAC addresses or an equivalent sort of identifier. I  
> wrote something for MAC plus something for things without a MAC. It  
> was looking less simple. There were problems with the multiple phones  
> on one MAC that added more complexity. Pretty soon, I was reinventing  
> UUID. UUID covered everything we needed and encompassed a huge set of  
> existing arguments on the subject. If you remove any part of UUID,  
> there is some use case that gets broken. It may be that the use case  
> is bogus but none the less the argument goes on.

Let me be clear, I'm not against UUID as such. In fact that would be a
perfectly valid spec to follow when implementing the registrar's
cookie-generation algorithm.

> Some devices store only  
> their MAC and the configuration system uses that to know what data to  
> hand down to them. I have seen some systems where the config handed  
> to the phone is based only on the which ethernet port in the switch  
> the device was connected to. Some residential ethernet to the home  
> systems do this so that if my neighbor plugs in his phone at my  
> house, it becomes my phone. In theory these systems would not need to  
> really store anything.

And that configuration handed down to them couldn't just include the
cookie?

> > And worst case is you need to garbage-collect some stale flows, in  
> > case
> > a UA happens to forget its ID, but that you need to be doing in any
> > case.
> >
> > We chose a similar approach over instance ID in PUBLISH, which uses
> > etags generated by the server to identify uniqueness of publisher
> > instances. With PUBLISH, UAs are also instructed to remember their
> > etags, even across reboots.
> 
> of course close to none of them do :-)

Touche! ;)

-- 
Aki Niemi
Nokia Research Center


_______________________________________________
Sip mailing list  https://www1.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [e-mail] for questions on current sip
Use [e-mail] for new developments on the application of sip