On Tue, Nov 13, 2012 at 1:02 AM, Noel Chiappa <j...@mercury.lcs.mit.edu>wrote:
> > However, I think the routing system has more problems 'coming down the > road' > (as the English expression goes) than just routing table explosion, and it > would be good to 'get out in front of them' (i.e. start working on > solutions > before they become incredibly painful). > Indeed. You don't configure nodes with locators (for all the very good reasons you > give). They are configured with identifiers. > Which is why it is identifiers which are configured, not locators. Those > are > _not_ changed when a site moves to a different ISP. By taking ILNP as an example, locators are subnet identifiers, a host is associated with multiple locators at multi-homing. This multiplicity of the locators, although the ID is single, implies multiple I-L vectors for a host and also affects DNS entries. This is equivalent to requiring IP hosts to be configured with multiple PA addresses. Have we solved the problem? Any improvement over the current IP situation? I'm afraid not. Multi-homing is not improved. -- DY
_______________________________________________ rrg mailing list rrg@irtf.org http://www.irtf.org/mailman/listinfo/rrg