> But since unless you manually or do some other finagling can't easily 
> establish trust if you don't have trust above, root-servers.net should only 
> sign after .net is signed at this point in the rollout.

The dependency on .net for the root name servers seems strange to me.

Intuitively, I should not have to trust .net to get a validated set of root 
name servers.

The names of the root name servers are somewhat arbitrary, and since they are 
very integral to the
root zone, it would seem more straight-forward to not put them into a public 
registry TLD, but rather
to use a special TLD ( e.g. "root-servers" or possibly a sub-domain of ARPA ). 
I don't see any
reason to use a sub-zone, the records may as well go in the root I think ( 
allows a secure resolver
to start up slightly faster ).

> And any PROPER useage of DNSSEC won't rely on root-servers.net ever being 
> signed at all, because its only on the name path for resolvers.

I think a proper use of DNSSEC is to use it to get a validated set of root 
server IP addresses.

This doesn't cure all the security ills of the world, but does constitute a 
small improvement in security,
especially for TLDs that have not yet been signed.

If TLDs also do not sign their name server domains, then a single blind spoof 
packet allows
an attacker to intercept all the traffic for a resolver. Even the root server 
traffic is somewhat
sensitive - it can often be what some end-user has just typed, which could well 
be confidential,
such as a password ( e.g. they think they are entering a password, but are 
actually typing into an address bar ).

Sure, it doesn't allow an attacker to forge signatures - but that's not the 
be-all and end-all of security.

I note that .se does sign it's name servers.
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to