El 14/05/2007, a las 11:59, Deguang Le escribió:
marcelo bagnulo braun schrieb:El 13/05/2007, a las 20:23, Deguang Le escribió:marcelo bagnulo braun schrieb:Hi Deguang, El 13/05/2007, a las 17:37, Deguang Le escribió:Dear Prof. Marcelo, page5: When H1 sends the first packet addressed to the CMULA of H2, the packet is intercepted and processed by the P- Shim6 box of the multihomed site.As above description, H1 sends the first packet, which uses the CMULA of H2 as the destination address. The CMULA is site-local adddress.rightSince the destination addresss of the first packet is the CMULA of H2 that belongs to H2's site, I wonder if this packet can be routed in H1's site.this packet is routed to the P-shim box in the H1 site. the P-Shim6 box of the H1 site will then establish the shim6 context whcih will have the CMULAs as ULIDs and regular PA addresses as locators. After that it will forward the packets outside H1 site towards H2 site containing the PA addresses as locators. So the packets will carry PA addresses in the IPv6 header once they leave the site. CMULAs are only routed inside the local site, so this would be compliant with proposed usage of ULAs.does this address your point?Thank you very much for your detailed explaination. :-)As you said, through the path from H1 to P-shim, the destination address of this packet is the CMULA of H2, which however is not routable in the H1's site (the reasions have been listed below). Normally, the CMULA of H2 can only be routable in the H2's site since it belongs to site-local address. right?right i see your point now.the idea is that the P-shim box will inject a route to the CMULA prefix, sinking all packets addressed to CMULA prefix (other than the local CMULA prefix)Thank you very much!Do you think whether your idea might violate the specification in RFC2374, namley "routers must not forward any packets with site-local source or destination addresses outside of the site."
i don't think so because CMULAs are routed inside the end sites, the source and the destination edn site, so no border router will havve to route them. Note that ULAs can be routed by other sites than the original one (consider the case of VPN for example)
So i don't see that this is a problem (but others may :-) what do you think? Regards, marcelo
With Best Regards, Deguangthen P-shim6 box would replace the CMULA to a globally routable PA address. So the CMULAs would be routed inside the sites and PA address will be routed in the interdomainthat makes sense? Regards, marceloTherefore, I only wonder how this packet could be routed from H1 to P-shim box correctly.With best regards, DeguangRegards, marceloThe reasions are ss follows:According to rfc2373, routers must not forward any packets with site-local source or destination addresses outside of the site.Reference [12] has the following description:Unique Local IPv6 Unicast Addresses are intended for local communications, usually inside of a site.RFC4193 provides operational guidelines that forbid default routing of local addresses between sites in section 4.1.By the way, the reference [12] seems have been updated by RFC 4193.[12] Hinden, R. and B. Haberman, "Centrally Assigned Unique Local IPv6 Unicast Addresses", draft-ietf-ipv6-ula-central-01 (workin progress), February 2005. Cheers, Deguang