Re: ETSI on Minimum Requirements for European ENUM Trials

[Jim Reid <]

>>>>> "Marco" == Marco Bernardi <marco.bernardi@localhost writes:

    Marco> All, I'm a little confused. I've got the impression we are
    Marco> discussing DNSSEC and a couple of other subjects (IPv6...)
    Marco> in relation to three different contexts: ETSI document on
    Marco> the minimum requirements, ENUM trial national
    Marco> implementation and finally ENUM commercial phase. This does
    Marco> not help the discussion.

On the contrary I think the discussion is helping to identify these
issues and where they figure in the overall scheme of things. The
three contexts you mention are inter-related. What's in or out of one
of them will have some impact the others.

    Marco> The aim of the ETSI document is to propose a MINIMUM set of
    Marco> requirements that can facilitate interoperability and
    Marco> interworking of ENUM national trials in Europe. The ETSI
    Marco> document is just about trial and interoperability and
    Marco> nothing more.

    Marco> My take is that DNSSEC is not a MINIMUM requirements for
    Marco> trial interoperability so it should not be recommended in
    Marco> the ETSI document (I would say it's the same for IPv6). The
    Marco> support of DNSSEC or other mechanisms to prevent spoofing
    Marco> is a national matter for each trial. The decision of
    Marco> supporting DNSSEC in the trial has also to take into
    Marco> account some practical aspects (e.g. costs, timeframe,
    Marco> extra complexity, etc...)  that are likely to be different
    Marco> from country to country. It would be interesting to have
    Marco> some trials (or portion of the same trial) with DNSSEC and
    Marco> others without and then compare the results/feedbacks

Indeed. But this means some minimal standards/requirements for those
trials that do involve DNSSEC so they can interoperate. And there
needs to be some convention for handling the interactions between
DNSSEC-aware and non-DNSSEC-aware entities. [ie If an application
expects a signed answer from the DNS and doesn't get that, what should
it do?] These are reasonable things for this ETSI document to cover:
not that I have any say on that of course since my employer is not an
ETSI member. Another topic the ETSI document could/should look at is
IPv6 interoperability: what if a name server or registry/registrar for
ENUM was IPv6 only?

    Marco> For the ENUM commercial phase I would say it's to early to
    Marco> take any final decision. Let's focus on the trials for the
    Marco> time being.

As far as I can tell, nobody is disagreeing with this. However ENUM
trials should look at as many aspects of a production ENUM service as
possible. DNSSEC is just one of those things. As is IPv6. And
EPP. And....