[arin-ppml] simple question about money

[michael.dillon at bt.com]

> > Because it is selling. 
> 
> So? I must confess I have little sympathy for this idea that "selling"
> is intrinsically evil. 

Who said it is evil? The fact is that selling IP addresses is not
allowed
by ARIN policy or by the contract that every ISP signs with ARIN. 

> I spent a lot of time studying China's 
> reform of its telecom sector in the 1990s, and in these RIR 
> debates I sometimes feel as if I am dealing with die hard CCP 
> members. If selling or trading addresses makes utilization 
> more efficient

Not really possible due to the way IPv4 addressing works. This
is a technical barrier that money cannot change.

> or increases access to v4 addresses by opening 
> up new resources that otherwise would be hoarded, then by all 
> means, let's embrace "selling."

We have opened up new resources with IPv6. Selling IPv4 only makes
it harder for people to access these new IPv6 resources.

> Or, as Deng Xiaoping said, I 
> don't care whether the cat is black or white as long as it 
> catches mice.

Typical CEO statement. IP addressing is not a problem that
will be solved with CEO culture. 

> The problem with your argument here is that it fails to take 
> into account the increasing scarcity, and the pressures that 
> will inevitably place upon ARIN and other RIRs. 

No, it is your argument that fails to take these pressures
into account. IPv6 exists today as a technology. Increasing
scarcity of IPv4 increases the pressure on companies to
deploy IPv6. When hardnosed business people look at the
possibility of buying IPv4 addresses to meet continuing needs
they quickly see that it is a losing proposition because the
increasing scarcity means that sellers will be few and prices
will be astronomical. Meanwhile, IPv6 technology exists today
and 80 percent of the v4-v6 interworking problems have already
been solved.

> I was going 
> to answer your argument in full but discovered that ARIN's 
> legal staff has already done so. Here is the analysis from 
> ARIN's counsel:
> 
> "No matter what policy ARIN implements, it seems likely that 
> there will be more disputes, and hence more legal risk, once 
> ARIN can no longer satisfy requests for v4 resources.  But if 
> ARIN attempted to continue its existing policy to prohibit 
> most transfers, counsel anticipates that widespread transfers 
> would nonetheless occur -- imposing significant future legal 
> costs including the costs of investigation, arbitration, and 
> litigation."

The key bit of this statement is "counsel anticipates". Basically
I don't believe that ARIN counsel is competent to forecast what
will happen since this is primarily a technical issue. Any company
that considers launching legal action over IPv4 addresses has to
weigh the negative impacts on the business of publicly admitting
that their business is not ready to cope with the new world of
IPv6, and that their growth prospects are therefore limited. 
I recommend forward thinking ISPs to keep their eye on the court
filings in Fairfax County, VA so that they know which company's
customers to target with their IPv6 sales force.

> Thank you, these are interesting empirical issues regarding 
> the degree to which addresses are fungible resources that 
> could be released and transferred. I would like to hear more 
> from ISPs about this. Or any other resrouces you know of that 
> clarify the issue.

May I suggest that you specifically look into the issues of
routability and the route filtering that most ISPs perform to
protect the stability and longevity of their core routers.

> so in that 
> respect it is like the IP address situation. It is cheaper 
> and easier to expand existing allocations than to abandon 
> completely the old space when massive legacy investments are 
> built around it.  

Interworking between IPv4 and IPv6 is possible. There are many
flavors of interworking, but it has been tested and it works
except for some corner cases. There are two years or so left
to work out the remaining technical issues so I think this is
being underrated as a solution. After all we are not talking 
about replacing IPv4 with IPv6. The reality is that we will have
a blended network for a generation, but IPv6 is where all the
growth and network expansion will occur. And that is because
it is easier and cheaper to build IPv6 networks than it is to
deploy double and triple NAT to make IPv4 feasible for growth.

--Michael Dillon