Hi,

Working on an effective solution for confidentiality
of stub to iterative resolvers seems an essential part
of the WG's mandate.  Input from P Vixie and others
seems to have this moving.

Thanks to Ted Hardie and Mark Andrews for expanding on
the issue of, and beginning a discussion of a potential
solution to, respectively, the iterative to authoritative
resolver DNS information leakage.

DNS information is clearly public information.  But that
does not mean that one needs to publish *who* is accessing
that public data.

As an ananlogy: the collection of books in a public library
may be considered 'public' information.  Nonetheless,
librarians expend effort to restrict access to the
information of who is borrowing the items.  i.e the
information itself is public, but the information of who
is accessing it does not need to be.

I encourage the WG to agree on a manner for confidentiality
between iterative and authoritative resolvers.  I suggest
that this would be an optional service (MAY not MUST),
at least to begin with.  The addition of encryption
adds a considerable weight; UDP -> TCP and handshakes
and ciphters.  Thus, it may take some time before many
authoritative resolvers are provisioned to provide
confidential query services.

Should the WG not wish to create the possibility of
confidentiality the resolvers, I hope that it at
least characterises the type and scope of what information
is being leaked.  Perhaps that should be done anyway.

I am quite happy to begin work on just that, should that
be desired.

Regards,  Hugo Connery
--
Technical University of Denmark

_______________________________________________
dns-privacy mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dns-privacy

Reply via email to