Hi, Wouldn't it be better to have ITRs store the Map-Servers (as defined in LISP ) instead of the mappings? This way mobility could be accounted for. Also NERD may be used in private scenarios. Then again, this is just a thought to improve NERD and Eliot just proposed to dismiss it.
Florin Robert Raszuk wrote: > Hi Robin, > > I very much agree with You and I find surprising the recommendation > from Eliot to give up work on it. > > IMHO we should not dismiss it. LISP-NERD is very much alike other > approaches where we store all the mappings in ITRs/ETRs/APRs depending > where we place such ITR/ETR/APR nodes. > > Sure perhaps requiring every CE to store all mappings is a stretch, > but we are told from every side that control plane memory is cheap. In > fact the scaling numbers presented by NERD were quite promising. > > Thx, > R. > >> Hi Eliot, >> >> Thanks for LISP-NERD. >> >> While there are good reasons for having an architecture in which >> every ITR does not need a full copy of the mapping information, I >> think NERD is notable for the simplicity of its architecture and >> directness of its operation: every ITR already knows the mapping, so >> there is no fussing around with looking up the mapping from a nearby >> or distant server. >> >> - Robin >> >> _______________________________________________ >> rrg mailing list >> [email protected] >> http://www.irtf.org/mailman/listinfo/rrg >> >> > > _______________________________________________ > rrg mailing list > [email protected] > http://www.irtf.org/mailman/listinfo/rrg _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
