> I basically agree with you re: QoS and market forces.  However none of
> what you have said addresses the concern of stability in the
> marketplace.  The other selection criteria that relates to stability is:
> 
> 7. Must submit contingency plan for support of its domains if business
> fails or decides to stop being DNA,
> 
> If the domains are going to "sit in the DNS", then what does this mean?

Peter may disagree, but in my opinion, it means "nothing".
It costs noone anything, and they get no service.  If they
need anything at all done (such as a redelegation), then
they'll need to find a new DNA to do it.

If we can separate the three independant functions of:

	o Host for the software/database
	o Authoritive name servers
	o DNA (the one who actually deals with clients)

then we also remove much of the problems in the event of a
DNA disappearing.

Authoritive name servers are redundant, so losing one is
bad, but can quickly be replaced without any real
disadvantage to the community (one less authority equals
a slightly slower lookup in 1/n cases).  We can have
organisations tender for the right to be an authority for
any 2LD, but the price will tend to zero, with the real
issue being a subjective recognition of connectivity.

The hoster of the software (let's call them a NIC) is more
of an issue.  If we require the the databases to be public,
it becomes less of a problem.  Even if a rogue disappears
with the database, there will be copies available elsewhere.

Losing a DNA is, in my opinion, a non event for the DNS.

MM