Lightly used IP addresses

Ken Chase ken at
Fri Aug 13 19:54:26 UTC 2010

On Fri, Aug 13, 2010 at 07:25:56PM +0000, Nathan Eisenberg said:

  >But the reality is that you asserted your intention to follow those
  guidelines when you requested the allocation, did you not?

  >If an upstream accepts announcements from a revoked block, what is to stop
  them from accepting announcements for an unallocated block?  I realize this
  precariously borders on committing a slippery slope fallacy, but I think
  it's a valid question to ask - a provider is either 'in compliance' with the
  guidelines, or 'not in compliance' with them.  Once you're 'not in
  compliance' a little bit, how can I have a valid trust relationship with you
  about the rest of it?

There's a difference - once the upstream is hooked on the revenue stream they're not
going to want to interfere with it. They might pass along some threats from ARIN
and/or levy their own, but I doubt they'd seriously make good on it and cut their
own hand off and lose the revenue.

That's for a deallocated block that was in good standing originally - this
assumes the contract for transit with the upstream included someone there
ensuring that the block was properly/legally allocated, and
WHOIS/RADB/Swip/yadda and everything else was properly notated and setup.
Going from good standing with a revenue stream for some months/years to bad is
different from accepting a bogon customer at the very start of the
arrangement. Lots of things would not lineup with a minimum of due dilligence,
and I suspect that most providers with any ethical slant will refuse to
provide service (scenario screams 'SPAMMER!' for one). That's alot different
from shutting off a revenue stream that was working well (sans spam) for
a year or more prior.

  >> following a corporation (yes, ARIN is a corporation) as if you were a sheep will
  >> empower them to do precisely this in the future.
  >There's no sheepism here.  The proposed situation represents a valid reason
  for revoking address space under the community developed guidelines.  I
  don't see the problem with following those guidelines, do you?

The reality is that following the guidelines is psychologically difficult in
harder times as we're experiencing now. Without any real repercussions for the
upstream for NOT cutting off the customer, balanced against the existing
revenue stream from the delinquent (assuming they're not delinquent with their
transit provider as well), it's not a hard calculation. I dont see much
'community' fallout occurring either, or we'd see it on this list. A few
transit providers have very poor reputations in the community (y'all know who
they are), and personally I won't purchase from them, but certainly none of
them have garnered this reputation by not cutting off ARIN delinquents. It's
just not publically available data - I dont think ARIN publishes this as I said,
and if they did I suspect it'd be a pretty busy-yet-boring mailing list (with
alot of screaming and name calling if it was open to public posting :).

  >> How many large carriers on this list would immediately halt announcing a
  >> downstream-in-good-financial-standing's prefixes just because ARIN say's
  >> they're delinquent?

  >That depends.  I vote with my wallet.  How many carriers want my business,
  and the business of other customers who (reasonably) expect compliance with
  the standing policies?  Do you want to do business with someone who's
  willing to break the rules everyone else is playing by?

IS everyone else playing by them? We dont really have data as I mentioned, or
I don't at least, so if anyone can provide stats (ARIN? some bulk numbers
without naming any names?) that'd be helpful in shaping this dicussion by
identifying how large the issue really is. Number of requests to upstreams to
halt announcements, and a mean and stddev on days-til-compliance for that
action (or how many delinquents were succesfully scared into paying ARIN by an
upstream's sternly worded warning would also be interesting). Unfortunately
such stats would also be good hard data for gamblers to model the risk/reward
profile on continuing to not pay. :) Shades of freakonomics game theory here...

  >Best Regards,
  >Nathan Eisenberg
  >Atlas Networks, LLC

Ken Chase - ken at - +1 416 897 6284 - Toronto CANADA
Heavy Computing - Clued bandwidth, colocation and managed linux VPS @151 Front St. W.

More information about the NANOG mailing list