On Saturday 17 July 2010 at 20:36:16 [email protected] sent: > Toni, > > > I think you are pursuing the wrong objective (which is global uniqueness of > the identifier).
Does anyone else think global/universal uniqueness of the identifier is the wrong objective? > > There are millions of people with the surname Li and yet they aren't a > serious problem for the postal services. For each letter sent to Li at a given postal address the sender already knows/expects the right Li is at that address. > IMHO ( I have expressed it many times and did also show how to get it done ) > the Internet routing should also become as smart so that local uniqueness > of any identifier will do. So, Heiner, the Internet routing system with its smartness would provide node identity for free roaming and session preservation? In my opinion the Internet has been successful so far because of IP addressing and the DNS, which are systems that give general uniqueness to nodes/interfaces. And I think these uniqueness systems give consistency to the sum of networks we have. > > Remember MADCAP. It was a similar mad idea (RFC2730). It tried to manage a > clash-free multicast address assignment - globally. It didn't fly at all. I may find the time to read RFC 2730. But would you summarize it here with regard to local/universal uniqeness of identifiers? > > Heiner > > Regards Toni > > > > > > > > -----Ursprüngliche Mitteilung----- > Von: Toni Stoev <[email protected]> > An: IRTF RRG <[email protected]> > Verschickt: Fr., 16. Jul. 2010, 16:29 > Thema: Re: [rrg] Universally unique identifier > > > Hi Klaas, > > On Friday 16 July 2010 at 15:57:06 Klaas Wierenga sent: > > On 7/16/10 9:11 AM, Toni Stoev wrote: > > > > Hi Toni, > > > > Interesting idea, however I am not conviced that a hierarchic model is > > desirable to guarantee uniqueness (I do understand the appeal, don't get > > me wrong). What concerns me is the fact that if you take delegation far > > enough you will end up with a very limited set of "N"-s per domain, and > > thus the ability to track users. > > If we don't set unwise limits, like the domain name system doesn't, we have > nothing to worry about. The market will regulate spanning. > > > Furthermore, I think the requirements for unique identifiers are not the > > same as for domain names. Arguably the delegation of subdomains is meant > > as a way to indicate some sort of subordinate relation, i.e. > > sales.acme.com is the sales department of acme etc. For uniqueness of > > identifiers I see no clear reason why you need a hierarchy, other than > > to avoid collissions. > > I agree, the reason differs here. Hierachy is a good way to control > distribution. > > > Don't you think it much easier to just use a method with a very low > > probability of collissions and deal with those? > > With online distribution we have flexiblity. One can change the identifier at > will. > > > Klaas > > Thanks > Toni > > > > A universally unique identifier is needed to let nodes roam > > > everywhere in the common network while preserving their ongoing > > > communication sessions. To keep identifier unique a system for > > > issuing, distribution and control over it is needed. The system has > > > to provide uniqueness and also be flexible and robust. > > > > > > There is an acting solution with much the said characteristics. That > > > is the DNS with its FQDN. The FQDN is by best effort universally > > > unique. And the DNS has proven itself to be surviving and moreover > > > prevailing. Because of node mobility the Dynamic DNS is the closer > > > match. With this technology a node can change its topological > > > location and retain a same identifier. But the FQDN is not quite > > > suitable for the hereby projected usage. It is made out of names that > > > are for human use. This not a networking best practice. > > > > > > So, can we turn the top supported network node identifier system into > > > a useful solution for also session identification supplement? And > > > how? > > > > > > Yes. By deciphering the N as "number". That's all to it. A dynamic > > > domain number system can supply the desired network node identifier. > > > The fully qualified domain number will be hierarchically structured, > > > fairly distributed, unique. > > > > > > Dear fellow researchers, your choice is welcome. > > > > > > Toni > _______________________________________________ > rrg mailing list > [email protected] > http://www.irtf.org/mailman/listinfo/rrg > > > _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
