In einer eMail vom 03.09.2009 23:20:15 Westeuropäische Sommerzeit schreibt [email protected]:
Heiner, My vision is that identity of nodes is to be based on relations with other nodes. Whereas my vision is that identity of nodes is to be based on relations to a neutral third and ever fix grid of "virtual/logical/abstract nodes". Otherwise you can map&encap to death. My impression: people don't want to abolish problems, they rather want to keep them as a permanent task to be mastered: scalability, looping... They can imagine Scoped Anycast, but not Scoped Anywhere of Unicast users.Here I cannot blame MIP4. They did an excellent job adopting a new requirement to a concept where roaming was not on the radar screen. LISP-ALT is a centralized care-of-address concept. What if both users roam as well as nodes? Potentially all users and all nodes? And even if you ignored the resulting new scalability and stretch problems you cannot even provide Scoped Anywhere service. However if a mobile node roamed within a particular geopatch no change of information is to be sent to the rest of the world, and also: you may provide the service to search for the destination in its "neighborhood" eventually without a home agent. By doing less you get much more... Especially, mutual knowledge of location within the general network. Maintaining this knowledge continuously and all over the network shall make node mobility natural and inconspicuous. I seek understanding and opinions on this vision, and I disfavor non-network based identity. without any deeper reasoning ? Heiner
_______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
