Hi Tony, > -----邮件原件----- > 发件人: Tony Li [mailto:[email protected]] > 发送时间: 2010年1月28日 15:33 > 收件人: Xu Xiaohu > 抄送: 'Xiaoliang Zhao'; 'RRG' > 主题: Re: [rrg] A concern with ILNP//re: critique of RANGI > > Xu Xiaohu wrote: > > > > > Binding transport associations to identifiers is a better choice. > > > In other words, using an ILNP host would be better than using a legacy > host. Yes, indeed. I agree with that point. > > > > However, > > like the home addresses in Mobile IP, the ILNP identifiers should also be > > routable in the Internet so as to allow the packets with destination being > > identifier to be forwarded to some home-agent like device. > > > You clearly missed the whole point: the identifier must be separated > from the locator, otherwise you have mangled semantics and you end up > back at LISP. This is a poor architecture. The whole point is to stop > compounding that original error.
Yes, it is ideal to separate them clearly from a clean-slate point of view. However, for incremental deployment purpose, I wonder whether the totally non-routable identifier is a practical choice. > > Then what's the > > business model for deploying such home-agent like devices based on the fact > > that the ILNP identifiers are flat. > > > There are no home-agent like devices in ILNP. Yes, I know. It's just an assumption. Best wishes, Xiaohu > > In addition, since the ILNP identifiers > > are not required to be globally unique, routing based on such identifiers > in > > Internet is almost impossible. > > > Exactly. > > Tony _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
