Re: Thinking differently about the site local problem (was: RE: site loc

On Mon, 31 Mar 2003 15:49:03 CST, Matt Crawford <crawdad(_at_)fnal(_dot_)gov>  
said:
> > Let's assume that there is a FooBar server in SiteA.  If another
> > node in SiteA (NodeA) is communicating via a multi-party application
> > to a node in SiteB (NodeB), and wants to refer NodeB to the FooBar
> > server in SiteA, what does it do?
> I thought we agreed, completely outside of IPv6 concerns, that
> shipping addresses in application data was bad. So NodeA refers
> NodeB to foobar-server.sitea.org. Q.E.F.
Yeah, we can agree all we want, but RFC959 still has a PORT command in it.

And until we've managed to move *all* the dain-bramaged applications to
Historical status, we're stuck with it.

And sometimes you have no *CHOICE* - if you're not shipping addresses around,
what *do* you put on a DNS A record?  This isn't facetiousness - it's a
real concern.  You can pass a hostname around instead of an address, and
when you look it up, you get back either a unique address (which you can
run with) or a site-local address (which you can't).  That's why RFC1918
has the prohibition against leaking private addresses into the DNS.

And let's face it guys - site-local is nothing but 1918 space on anabolic
steroids.  You thought it was hard to handle now, wait till it comes back
with a full blown case of "roid rage"....

pgpzhYLHl4jKb.pgp
Description: PGP signature