[arin-ppml] Revised -- Policy Proposal 2009-4: IPv4 RecoveryFund

[Ted Mittelstaedt]

> -----Original Message-----
> From: David Farmer [mailto:farmer at umn.edu] 
> Sent: Monday, April 13, 2009 3:54 PM
> To: Ted Mittelstaedt; arin-ppml at arin.net
> Subject: RE: [arin-ppml] Revised -- Policy Proposal 2009-4: 
> IPv4 RecoveryFund
> 
> On 13 Apr 2009 Ted Mittelstaedt wrote:
> 
> > > I disagree, publicizing that it is expensive and risky to 
> stick with
> > > IPv4 would help drive people toward the stability and generally 
> > > lower cost of IPv6.
> > > 
> > 
> > NOT publicizing it is even more effective, as people 
> invariably assume 
> > it's more expensive than it really is.
> > 
> > Which do you immediately assume is more expensive?
> > 
> > An expensive
> > snooty restaurant with a 5 star cook that charges $60 a steak?
> > Or an expensive snooty restaurant with a 5 star cook that 
> doesn't put 
> > the prices on the menu and has the attitude that if you 
> have to ask, 
> > you cannot afford it.
> 
> Ok, that works for some people, but, what if you have the 
> boss that says prove to them it will be more expensive to 
> stay with IPv4, and cheaper to go with IPv6.  Without public 
> data, on how expensive IPv4 is, how do you make a business case that
> IPv6 is the better way to go? 
> 

That's a boss that isn't interested in going to IPv6 and you
ought to know that one by now.  In other words, the admin may
make a very good case for going to IPv6 - but that boss will
merely find something he doesn't like about it and focus
on that.

If your boss can't be convinced to go to IPv6 by the argument
that IPv4 runout is happening on June 19th 2011 and it will
affect every ISP in the world, so we need to get going on
IPv6 now, then they aren't going to be impressed until year
2012 when a major customer calls up and says "I need IPv6
yesterday, you don't have it? OK I'm going to disconnect now
and go down the street to your competitor"

And even then, there's bosses out there who will say we don't
need to go to IPv6 until more than 10% of our customers have
left to go to a competitor.

Some bosses simply won't be convinced they need to go to IPv6
until the place is burning down because you haven't gone to
it, and then they will want it yesterday.

> I agree that the current daily or hourly price could be use 
> in ways the community doesn't want.  I can accept that there 
> should be a 15 or 30 day delay on the release of pricing 
> data, so maybe you don't release February's data until March 
> 15th or March 31st.  But, we need reasonably current pricing 
> data for people to make business cases with.  Even the 
> business case we would like them to make, like going to IPv6, 
> will take some data. 

The fundamental problem with a business case to going to IPv6
is that the benefit to doing it is effectively zero until
the content providers on the Internet that your customers want
to reach (or the customers that your content-providing customers
want to reach) are all on IPv6 and are beginning to wind down
their IPv4 operations.  Then, the benefit shoots immediately
to 1000%.  There is no soft-uptake.  So the admin who wants to
do it is left arguing that the benefit of doing it now is less
work to do it later on, but that's a very speculative and hard
to quantify argument.

I suspect a lot of admins out there simply take the
approach that I took at my company which is to just start working
on it by selecting IPv6-compliant gear, without even giving the
boss a say in it.  Then eventually when everything can handle it,
switching over to it.

> The trick is to provide data, without 
> it being so current that the release of the data might cause 
> speculative blips in the pricing, there has got to be a 
> balance some where.
> 

I agree with this approach.

Ted