Excerpts from William Herrin on Fri, Apr 03, 2009 01:04:26PM -0400: > On Fri, Apr 3, 2009 at 11:21 AM, Scott Brim <[email protected]> > wrote: > > A locator names a point of attachment at a given layer. If an > > endpoint changes its points of attachment at that layer, > > associations between the endpoint and locators will change. > > Hi Scott, > > I like this a lot better. > > It's still missing something though. LISP's LOC is definately a kind > of locator but it doesn't change when I move one of my servers from > one internal LAN to another internal LAN and assign it an address > from a different subnet. I've changed my layer-3 attachment but the > locator, which is also layer-3, doesn't change. > > Maybe something more like: > > A locator names a point of attachment within the scope of the > network topology for which the locator is used. If an endpoint > changes its points of attachment within that scope, associations > between the endpoint and locators will change.
I'm having trouble deciding what that means, and I'm enthusiastically trying, so I'm sure others will have even more trouble. :-p A LISP RLOC names a point of attachment point for an ETR, and IP forwarding in the core uses it as a locator to determine how to forward the packet to that PoA. Once the packet is delivered, the ETR strips off the LISP header, exposes the header underneath, and uses the destination address in that as a locator (which names an interface on the destination). That all fits within the definition at the top. There are two "endpoints" in the sense of the definition: the ETR and the ultimate destination. LISP says: "Routing Locator (RLOC): the IPv4 or IPv6 address of an egress tunnel router (ETR). It is the output of a EID-to-RLOC mapping lookup. An EID maps to one or more RLOCs." It doesn't say that an RLOC names a point of attachment for the ultimate endpoint, just for delivery of the LISP-encapsulated packet. IMHO the association between an RLOC and an ultimate delivery point of attachment is in the LISP encaps/decaps functions, and doesn't have to be covered by the definition. OK? Scott _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
