Hi Klaas, On Thursday 22 July 2010 at 13:05:45 Klaas Wierenga sent: > On 7/22/10 11:33 AM, Toni Stoev wrote: > > Toni, > > > On Wednesday 21 July 2010 at 18:23:32 Toni Stoev sent: > >> On Tuesday 20 July 2010 at 12:36:56 Javier Ubillos sent: > >>> In the case where the service would be provided by random sites > >>> on the Internet, well, it is already provided for free today, but > >>> I don't know how it would scale (or how a business model would > >>> work) if all hosts start registering names automatically. > >> > >> Many non-mobile service ISPs employ Dynamic DNS too. For inter-ISP > >> roaming random/devoted sites are the solution. The model is quite > >> analogous to the one with names. > > > > On Thursday 22 July 2010 at 00:19:31 Scott Weeks sent: > >> Operators in the commercial area have to satisfy their customers > >> and the lack of privacy will cause concern among customers. It's > >> already doing so or TOR and things like it wouldn't exist. I > >> believe this concern will only increase and not decrease over time. > >> Thus the operators will care as it all drives the bottom line. > > > > Privacy depends on operators' loyalty. But having a choice of > > identifier providers, we can rely on identity privacy, even having > > only one connectivity provider. > > I don't think the latter is very realistic. On the application layer, > sure, you can have any number of identity providers. But I don't think > it is likely that you will have a supplier of a network identifier that > is not your connectivity provider.
It would not be compulsory to have one, but if you would have the possibility to choose a remote network identity provider, your privacy would have more of a chance. And choice is the advantage of the popular dynamic DNS service providers. > Sure, I can imagine a world in which > TOR like companies will emerge that provide some sort of proxy and that > promise to keep your identity secret, but if you want to do that any any > significant scale you will run into problems. Let's not create a design > that forces us 10 years down the road to solve the GUPI-PII (global > unique persistent identifier - person identifiable information) > problem...... I am all for incremental design, but having a tight > coupling between humans and persistent identifiers is asking for trouble > imo. I agree. Toni > Klaas > _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
