Le 29 juil. 2011 à 16:27, Cameron Byrne a écrit : > > On Jul 29, 2011 7:15 AM, "Rémi Després" <[email protected]> wrote: > > > > > > Le 29 juil. 2011 à 15:48, Satoru Matsushima a écrit : > > > > > Cameron, > > ... > > > Stateless is not exclusive with stateful one, and also NAT64. > > > I agree with Hui. > > > > +1 > > RD > > > > Ok. I do think the draft would benefit from narrowing some of the options. It > seemed to me that statelessness was a key feature, but I may have misread. > Some providers may indeed prefer to have only a stateless solution, especially in for DSL networks where ISP's provide CPE's, but there is no technical reason preventing both solutions to coexist. > ... I am skeptical about stateless operation scaling. I don't even have 2000 > ports per user (growth to 2015 ...), > Since concerned ports are only those that are residually used in IPv4 (while more and more traffic should sooner or later become IPv6), 2000 IPv4 ports per user seems quite sufficient. If not, would you have figures that show it?
(Besides, note that 16 users per IPv4 address (a quite significant increase), 3840 ports are available per customer (almost twice 2000). > and I don't want to get into a position where I have to tactically move ipv4 > addresses around to meet demand. > If each 3GPP user gets a different IPv4 address each time it is connected, this shouldn't be a concern. is there a different use case you have in mind? Last point: are you in fact proposing that 3GPP IPv4v6 bearers be _replaced_ by IPv6-only bearers with IPv4/IPv6 translation, or that both should coexist to widen the number of options? Regards, RD > > > > >
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
