[Sipping] Re: P2P why not DNS for discovery?

[Juha Heinanen <jh at lohi.tutpro.com>]

the problem that i see with p2p solutions like skype is that they
introduce a new identity for me, which is not the same or related to my
email address jh at tutpro.com.  whatever the p2p sip solution is i must be
able to keep on using sip:jh at tutpro.com as my sip identity so that
people who know my email address can guess my sip uri.

this is possible with the dns based discovery mechanism i proposed
earlier only if the user can update sip srv record of user.host, which
is not possible, for example, in case the user's email address is
foo at hotmail.com, because bill gates may not like the idea.

in order to overcome this shortcoming, we could do the following:

1) register for p2p discovery purposes a new domain, e.g., p2p.arpa.

2) p2p.arpa accepts delegation request for subdomain user.host.p2p.arpa
   if it gets a positive reply to confirmation email it sends to
   user at host.

3) once delegation is done and name server of user.host.p2p.arpa zone is
   running, user can start dynamically updating sip srv records of
   user.host.p2p.arpa.

4) when a p2p sip UA wants to call sip:user at host, it sends sip svr
   request for domain user.host.p2p.arpa.

this way anyone with email address from any domain can gets its sip srv
records updated.  the mechanism also provides some assurance that calls
to sip:user at host actually reach the same person whose email address is
user at host.

a trial implementation of the above that does everything automatically
(including setting up of dns server zone user.host.p2p.arpa) can be done
in a week.

-- juha

_______________________________________________
Sipping mailing list  https://www1.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors at cs.columbia.edu for questions on current sip
Use sip at ietf.org for new developments of core SIP