[ppml] Soliciting comments: IPv4 to IPv6 fast migration

[Ted Mittelstaedt]

>-----Original Message-----
>From: ppml-bounces at arin.net [mailto:ppml-bounces at arin.net]On Behalf Of
>William Herrin
>Sent: Tuesday, July 24, 2007 12:28 PM
>To: ARIN Address Policy
>Subject: [ppml] Soliciting comments: IPv4 to IPv6 fast migration
>
>
>Hi Folks,
>
>Following your comments and some helpful off-list discussion, I've
>prepared a replacement for last month's "IPv4 to IPv6 Migration
>Incentive Address Space" proposal. With some mild tweaks to the
>existing 6to4 protocol, it seeks to address four problems ARIN faces:
>
>1. The looming exhaustion of the IPv4 space.
>2. Obsolete and incorrect legacy IPv4 registration and contact information.
>3. Legacy IPv4 registrants don't pay their fair share.

You know, William,

  I and many others have raised this payment issue repeatedly.  The legacy
holders appear to generally not be even on this list at all or even paying
attention, or they are getting apologists to make the lame argument that
the RIR's tracking these people benefits me more than them, so as a paying
member I must want to pay for them to be tracked.  Of course, this argument
completely ignores that I have already asked to NOT pay for them to be
tracked and I am not interested in being told by the apologists what is in
my best interest.

  Frankly, I think it is a lost cause.  We all know that IPv4 is going to be
around for many more years, and the legacy holders are going to continue to
use
their IPv4 allocations for many more years free of charge.  Furthermore as
recent
discussion seems to indicate that a lot of people, not just legacy holders,
are vehemently
opposed to derouting IPv4 once IPv6 becomes in wide usage on the Internet,
it
is very likely that 10 or 20 years post IPv4-runout (long after both of us
are retired no
doubt) the Internet will STILL have both IPv4 and IPv6 on it, with NO end to
the
IPv4 in sight.

  The dirty little secret that it appears a lot of people want to cover up
is
that ONLY organizations with EXPANDING addressing needs are going to need to
bother to migrate to IPv6, immediately post-IPv4-runout.  It is going to
take
many years before must-have servers and must-connect locations on the
Internet
appear that are ONLY IPv6.  And before that happens, huge incentive will
exist
for new orgs and new deployments to be available to the IPv4 world - because
nobody is going to know which one of their potential customers is still on
an
IPv4-only connection.  Clearly, the idea that runout is going to spur IPv6
switchover
is bankrupt.

  I will also point out that NO financial incentive exists for orgs who HAVE
completely
switched to IPv6 to release their IPv4 holdings, EVEN IF the ENTIRE INTERNET
has switched to IPv6.  Only if it is a NEW org that has NEVER been touched
by IPv4 -
never configured an IPv4 address on anything - then incentive exists to NOT
use IPv4 on an IPv6 Internet.  But, if the org has gear that once upon a
time
ran IPv4, it is not a financial incentive to stop running it.

  The situation is very similar to networks that ran TCP/IP and IPX.  When
NetWare
NCP became available on TCP/IP these orgs didn't just shut off IPX.  Same
goes for
orgs that ran Apple's Ethertalk in conjunction with TCP/IP.  They continued
to run it
concurrently and only as brand new gear never sullied with IPX or Ethertalk
was
deployed did they stop running it.  And until they shut it off at their
servers,
people running the old Mac's without TCP/IP, or the old DOS-based ipx.com
clients,
they didn't have to switch.  Granted, small scale, but similar.

  I call your attention to the US public switched telephone network,
instituted
over a century ago.  POTS line voltage: - 48v, a century old standard still
in use.
Rotary dialing - Strowger switching, 1891, a century old standard still in
use.
DTMF, late 50's a half-century standard still in use.

  The Internet - IPv4, late 70's early 80's - a 25 year old standard still
in use -
and if you don't think it's not going to make it another 15 years
post-runout, your
out of your mind.

  So don't hope for the legacy IPv4 holders to be paying anything, anytime
soon.

  Maybe what is needed is a policy proposal that has as it's only line item
a
NON-binding demand for all legacy holders to immediately sign RSA's for all
their legacy holdings and start paying the fees.  With no penalties if they
do
not.  In that way we might move the fee discussion off the RIRs and to the
legacy
holders where it belongs.

  In other words, if the stated policy of the RIR is that all legacy holders
are
to sign RSA's for all their holdings and be subject to addressing fees, then
the
discussion will become one of which legacy holders are choosing to do the
right thing
and pay the fees, and which are not, and legacy holder fee payment will no
longer be
a line item that will come up in policy proposal.  For legacy holders that
choose to
ignore the demand, no penalties will follow - which means they will be
unable to file
lawsuits to overturn anything, of course, since no damages will have been
suffered - but
the moral weight of being classified as a deadbeat or non-participant might
have meaning
in policy discussions.

  And, if such a policy proposal was voted down it would also help to end
the fee
discussion.  In other words - we put together a policy proposal telling the
legacy
holders they have to sign an RSA - but if they don't, we won't punish them -
and we
couldn't even get that voted in.  In other words a completely toothless fee
proposal
couldn't even get passed.  Kind of like the scene from Animal House; "thank
you sir,
may I have another"

Ted