[Secdispatch] GNS at dinrg

[Melinda Shore <[e-mail]>]

I had a conflict this morning and wasn't able to participate
in the secdispatch session, but I've been asked about the GNS
presentation at dinrg at IETF 104 (dinrg is an IRTF research
group investigating decentralizing internet infrastructure
elements).

The question of adopting GNS as an RG deliverable never came
up.  Research groups operate differently from IETF working groups
and are not heavily driven by deliverables.  In general, the
feedback from the floor expressed concern about scalability and
key management, DHT performance, and (as always), governance.
I do think, however, that GNS remains more of a research project
than a standardizable technology, as one of the bottom-line
questions for the standardization of pretty much anything is
whether or not it's deployable on the public internet.  Even
if the technical questions about GNS are resolvable (little
DNS joke in there) the governance question looks intractible
("we would advise ICANN not to assign any more names" doesn't
seem as if something ICANN would be likely to act on).

Also, it's become more common to talk about DNS as part of
the internet control plane and we're seeing increased use of
the DNS to bind things other than addresses to domain names.
That has implications for performance and scalability (see
above), as well as the security model.  Perhaps that was
addressed in today's presentation - I'll watch it once the
video is posted.  It's interesting technology but unless
they can both work out a very incremental deployment model
and deal with the governance question (and I expect those two
things can be linked) I'm not that optimistic about its
deployability as a core piece of internet infrastructure.
(I'd love to be wrong about that!).

Melinda

-- 
Melinda Shore
[e-mail]

Software longa, hardware brevis