Agree Noel. You must have a mapping as well. If you don't have a mapping registered and there is no route in the underlying, then the EID is not usable.
If both are in each system, then it depends on ITR behavior which is used. Dino On Oct 31, 2013, at 1:08 PM, [email protected] (Noel Chiappa) wrote: >> From: Dino Farinacci <[email protected]> > >> We are not saying this entire block is being used for global deployment >> of LISP. And no one is saying LISP can't succeed without this block. >> ... >> It does not mean that the current forwarding paradigm does not work >> or cannot work. >> We must not and will not require reassignment of addresses to use >> LISP. ... existing allocations and futures allocations can be EIDs. > > These are all _really_ excellent points, and the document should make them > clearly, and at the top. (Sorry if it already does this - haven't read any > recent versions, too busy.) And also the point about how the special prefix > is not going to be the way we determine whether an address is an EID. > >> An address becomes an EID when it no longer is advertised by an edge BGP >> router (from a tail or stub portion of the Internet topology). > > Minor quibble, because I prefer the 'is there a mapping available' as the > 'gold standard' test for EID-ness. > > For backward compatability with 'legacy' hosts, many EIDs _are_ advertized > into the global routing (either by the ETR, or a PITR, etc), so 'in/not-in > the global routing table' doesn't tell us much about EID-ness. > > Noel > _______________________________________________ > lisp mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/lisp _______________________________________________ lisp mailing list [email protected] https://www.ietf.org/mailman/listinfo/lisp
