[gnso-rds-pdp-wg] Now open: 18 January Poll on Purpose

[Andrew Sullivan]

On Thu, Jan 26, 2017 at 04:08:05PM -0000, Rob Golding wrote:

> Indeed, in fact as the resolver method is more reliable, more
> accurate, faster, necessary and so on - why are we duplicating it in
> WHOIS at all ?!?

Because it's not duplicate data.  The RDS tells you what the
registration system thinks, not what the DNS thinks.  The two systems
_ought_ to have the same data, but that can only be checked by looking
at both.  (This is also, note, the reason that a "centralized" system
that holds all the data for all registries is as astonishingly bad
idea, because it creates yet a new data sync problem that cannot be
checked.)

When things break, one thing any competent network admin does is check
the DNS to make sure something isn't wrong.  If that doesn't seem
broken, one immediately checks whois/RDDS to see whether what's in the
DNS is what's _supposed_ to be there.  I admit to being a newbie to
the Internet, since I didn't join it until some time in the 1990s, but
as near as I can tell this is what people have _always_ done to
diagonose problems.  A distributed database system with a lot of
caching benefits from an independent way to check that it is working
correctly, so that diagnosis from the far-flung edges of the network
need not require everyone checking with a central authority all the
time.  Having an RDS is useful for making that work reliably.  The
Internet has been the success it is because of this distributed
nature, and I think it is jaw-droppingly ridiculous that we continue
to debate whether that is something we want to maintain.

There are lots of other uses of the current whois system that I think
are bogus (I think, for instance, that the encroachment of
intellectual property claims on the DNS has been an unmitigated
disaster for the Internet).  But this technical use is the basic point
of the RDS facility, and I think it is plainly useful.

Best regards,

A

-- 
Andrew Sullivan
ajs at anvilwalrusden.com