Moin,
> SHOULD is a cop out.  

Given that this is already an implemented feature in Unbound, I'd argue
that a <bcp14>MUST NOT</bcp14> might not be ideal here.

> What we are designing SHOULD work everywhere WITHOUT
> manual intervention.

I would have to manually intervene a lot more by setting up CLATD on my
DNS recursor instead of just dropping two lines of configuration into
the config.


> Translating IP addresses in a DNS server is unnecessary work.

I would argue that "Get A, move bits, open v6 socket, make query" is a
lot less work than "Get A, open v4 socket, kernel takes packet does
CLAT, IPv6 packet goes out, send query", especially given the state of
CLAT on servers, and especially-especially given that CLAT regularly
involves more state on host OSes due to the 'magic' needed to redirect
packets to the CLAT.

> It is also a security flaw as you are potentially redirecting DNS
> queries to places other than where they are intended to go according
> to the routing table.

I'd say that this is a bit far fetched. Also, on Linux, CLAT is usually
userspace; So you would already have to have exceptions w.r.t. what
should or should not be going through CLAT and what not.


> > The ways to configure one's own network are endless. What if there
> > are two interfaces each with their own PREF64? What if the second
> > Interface only offers reachability for RFC1918? What if the
> > operator has been squatting non-RFC1918 space and the 2.0.0.0/24
> > reachable via Interface B is _not_ the one they want to use for DNS
> > resolution?
> 
> All the more reason to NOT do this.  It’s not something that can be
> done safely.

But that was essentially you use-case? Or did I miss sth. there?

With best regards,
Tobias


-- 
Dr.-Ing. Tobias Fiebig
T +31 616 80 98 99
M [email protected]
Pronouns: he/him/his

_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to