And if we are talking in your context then a link-local multicast packet (or unicast for that matter) should STAY ON LINK, which means it should not be layer-3 encapsulated. So we are in agreement.
Note, STAY ON LINK, in most circles, means contained within a subnet/VLAN/layer-2 domain. Dino On Mar 24, 2013, at 7:24 PM, "Joel M. Halpern" <[email protected]> wrote: > I guess I saw "mobility" and thought it was abut the stuff we have seen in > drafts abut LISP for mobile devices (cellular and WiFi). If the discussion > topic is VM movement and L2 extension, yes, I agree with what you have said. > > Yours, > Joel > > On 3/24/2013 8:48 PM, Dino Farinacci wrote: >> Joel - if we talk about link-local multicast being forwarded it is in >> context of an L2 overlay where EIDs are MAC addresses. So NVO3 related work, >> even though NVO3 is not limited to L2 overlays. >> >> Dino >> >> On Mar 24, 2013, at 3:54 PM, "Joel M. Halpern" <[email protected]> wrote: >> >>> Maybe I am naive, but it strikes me that trying to extend link-local IPv6 >>> multicasts with LISP is probably a bad idea. Within the mobile device >>> scope, the registration mechanisms would seem to provide better tools for >>> things like address duplication prevention. >>> >>> While I don't know if the documents capture it properly, the fragmentation >>> effect of mobile devices on EID resolution blocks is such that I generally >>> would want to keep them separate from the fixed devices. >>> >>> Yours, >>> Joel >>> >>> On 3/24/2013 6:13 PM, Paul Vinciguerra wrote: >>>> I know I’m showing up late to the party, but can someone expand upon >>>> this from the RFC? >>>> >>>> *5. Source Addresses versus Group Addresses* >>>> >>>> Multicast group addresses don't have to be associated with either the >>>> >>>> EID or RLOC namespace. They actually are a namespace of their own >>>> >>>> that can be treated as logical with relatively opaque allocation. >>>> >>>> So, by their nature, they don't detract from an incremental >>>> >>>> deployment of LISP-Multicast. >>>> >>>> The documents make reference to: >>>> >>>> (S-EID,G) >>>> >>>> (S-RLOC,G) >>>> >>>> It seems that while both these may be numerically identical (eg. >>>> 239.255.255.254), in terms of LISP, doesn’t (S,G) have to much more >>>> complex and represent either (S-EID,G-EID(IID)) and (S-RLOC,G-RLOC(IID))? >>>> >>>> Once we introduce LISP mobility, what does (G) actually mean, especially >>>> when (G) can refer to link-local multicast for a host that has roamed >>>> off of its home subnet? >>>> >>>> Paul >>>> >>>> >>>> >>>> _______________________________________________ >>>> lisp mailing list >>>> [email protected] >>>> https://www.ietf.org/mailman/listinfo/lisp >>> _______________________________________________ >>> lisp mailing list >>> [email protected] >>> https://www.ietf.org/mailman/listinfo/lisp >> _______________________________________________ lisp mailing list [email protected] https://www.ietf.org/mailman/listinfo/lisp
