Breaking up a class for delegation

[Niall O'Reilly]

On 26 Sep 2007, at 00:26, Curt Sampson wrote:

> It's been a proposed standard since 1999, I see.
> How widespread is DNAME support in resolvers?

	It's not clear to me whether you mean 'full-function' or 'stub'.

	Either way, I'm not sure.  The synthesized-on-the-fly CNAME tends
	to mask any deficiency in this area.

	I do know that BIND has had it for quite a while.

	I'm led to believe that a resolver can't properly support DNSSEC
	unless it supports DNAME.  I haven't fully understood the argument,
	but understand that some unacceptable corner cases arise otherwise.

	/Niall