North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: SWIP update intervals (was: Re: Getting an AS and /18)

  • From: Eric Brunner-Williams in Portland Maine
  • Date: Tue Jul 03 21:29:41 2001

...
> because the major use of the whois data has become spam

I'll ask my friends in the DMA (Direct Marketing Association) and the IAB
(the _other_ IAB, the one that's Got Milk, the Internet Advertizing Board)
about the quality of the whois data from their perspective. My sense is it
(registrant email/smail/isp data) is cheaper than $.10/unit mailing lists,
and not as well maintained as $1/unit lists. I just can't recall a major
on-line add campaign that spammed.

I suspect that whois spammers are fairly far down in the food chain, since
it is "free data", and utterly unmaintained historically.

The interesting question (to me) is if "social data" [1] is policied, a la
p3p-esque metadata describing the data collection practices at the point of
collection, and subsequent onward-transport (reseller -> registrar (and in
the case of a "thick registry" [2]) -> registry), then can (r)whois servers
policy queries and meet the operational requirements of all of the (r)whois
communities?

Generallized, if metadata is available to policy data, is this sufficient
mechanism to jurisdictionalize (j19n) the privacy (US) and data collection
(OEDC & EU) policy regimes? I like this approach, but my year-in-P3P may
have affected my vision. I'm putting it in XRP (provreg WG, Apps AD), as
an option for registry operators.

Another approach to the problem is simply to assert that registrant contact
information serves no useful operational purpose, and serve only technical
and administrative contact data via whois:43, and declare that as meeting
the operational requirements of all ... and putting the trademark et al on
a distinct port, with a _vastly_ better query mechanism, and providing some
access control mechanism(s) for "sponsored interfaces" to the registry data.
This should a) satisfy operators, and b) satisfy trademark owners, and c)
leave spammers out in the cold.

If you use whois:43 in an innovative way, please drop me a line, with the
string "whoisFIX" in the subject line. There will be another whois BOF at
London.

Thanks for the opening line Randy. Who was that masked man?

Eric

[1] draft-ietf-provreg-epp-contact-02.txt
[2] draft-ietf-provreg-dn-defn-01.txt