Thus spake "S Woodside" <sbwoodside@yahoo.com> > Then if you conclude that policy domains are a Good Thing, or at least > Necessary Evil, then why is there all this talk to design a network > that can somehow route around them? There's reasonable arguments against private addresses, but unless allocation policies are radically different than IPv4 practice, expecting a public address for every host is a pipe dream. Sidebar: a thread has recently popped up on nanog regarding the practice of assigning public addresses to unconnected networks or hosts behind firewalls. It's not clear whether those for or against are in the majority, but the mere presence of the debate is rather telling. > My point is that A sends B a third-party address C, and the policy of > the domain is "you can't route that outside my domain" then it doesn't > matter whether C is site local, global, uses DNS, or whatever. Policy > says it still won't route! The expected usage model was that all hosts would have a site-local address as well as zero to many global addresses. Since we don't yet know how to handle multiple addresses per host cleanly, the removal of site-locals is thought to reduce the problem's complexity since site-locals are for some reason assumed to have different semantics than globals. S Stephen Sprunk "God does not play dice." --Albert Einstein CCIE #3723 "God is an inveterate gambler, and He throws the K5SSS dice at every possible opportunity." --Stephen Hawking