Peng, You are right. The 4rd is very similar to stateless 4over6. If the v6 routing is enabled between 4to6 clients, we don't need to use the 4over6 BR. The main difference is IPv4 address sharing. Perhaps, we can consider to work together for this.
/Yiu From: Peng Wu <[email protected]<mailto:[email protected]>> Date: Wed, 23 Mar 2011 23:30:18 +0800 To: R閙i_Despr閟 <[email protected]<mailto:[email protected]>> Cc: <Lee>, Microsoft Office User <[email protected]<mailto:[email protected]>>, Shishio Tsuchiya <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Re: Re: [Softwires] WG Review: Recharter of Softwires (softwire) I think I've got what you guys mean. It's the same problem we're talking about. And, as long as it's stateless IPv4-IPv6 address mapping with routing support, traffic between CEs naturally doesn't need to go through the concentrator. As to address sharing, I guess the basic idea is leveraging some bits in port space and some bits in IPv6 address, right? Again I think it's one step further upon stateless IPv4-over-IPv6. I'll read the draft to get details. 2011-03-23 ________________________________ Peng Wu PhD candidate Department of Computer Science & Technology Tsinghua University, Beijing, China ________________________________ From： R閙i_Despr閟 Date： 2011-03-23 23:15:10 To： Peng Wu CC： Lee, Yiu; Shishio Tsuchiya; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]> Subject： Re: [Softwires] WG Review: Recharter of Softwires (softwire) Hello Peng, You are right, there is some commonality of objective. There are however important differences: - 6over4 is hub-and-spoke => It might be envisaged as a DS-lite extension. This isn't the case of 4rd which has direct paths between CE's. - 6over4 as it is deals only with mappings of addresses. 4rd as it is includes a solution for shared IPv4 addresses (restricted port sets are algorithmically derived from bits of IPv6 prefixes). This could be added to 6over4 to extend its scope. In view of the commonality, your comments on 4rd will be most welcome. Regards, RD Le 23 mars 2011 à 15:43, Peng Wu a écrit : > Hi guys, > > I haven't cover the 4rd draft yet, but are you talking about the stateless > case of host 4over6？ Maybe the two mechanism have overlaps. > > The basic idea is to coupling IPv4 and IPv6 address of hosts/cpes using > prefix mapping, in order to achieve stateless encapsulation on data > forwarding. > > It's an ongoing work and we've presented this draft in IETF before. > > Please take a glance at section 5 of the 02 version draft > > http://tools.ietf.org/id/draft-cui-softwire-host-4over6-02.txt > > > 2011-03-23 > Peng Wu > PhD candidate > Department of Computer Science & Technology > Tsinghua University, Beijing, China > From： Lee, Yiu > Date： 2011-03-23 22:06:11 > To： Shishio Tsuchiya; [email protected]<mailto:[email protected]> > CC： [email protected]<mailto:[email protected]>; > [email protected]<mailto:[email protected]>; > [email protected]<mailto:[email protected]>; > [email protected]<mailto:[email protected]> > Subject： Re: [Softwires] WG Review: Recharter of Softwires (softwire) > +1 > I think this is an interesting solution and the WG should spend more time > to discuss this. > /Yiu
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
