[ppml] Incentive to legacy address holders

[John Curran]

At 7:17 PM -0500 7/8/07, Robert Bonomi wrote:
>The existing system works *ONLY* because of 'voluntary co-operation', because
>'enlightened self-interest' indicates, *presently*,  that cooperation with
>those agencies is desirable.

You are correct;  the fact is that the Internet was built on
voluntary cooperation, and making things work together
was the entire reason that there was an IANA (Jon) to
coordinate IP address assignments. 

In general, there's been fairly good cooperation among ISPs
over the decades...  we've managed to do lots of things (like
DNS, CIDR, bogon filtering, MD5 for BGP, anycast DNS, etc.)

We've also had some areas that have been more challenging
(e.g. route registries, prefix filtering policies, any secure BGP,
DNSSEC, routing table containment)

Decision making based on distributed enlightened self-interest
is what we have to work with, but let's also recognize that it's
remarkable bad at handling situations that require significant
coordinated efforts well in advance of any imminent crisis...

>Now, if/when the time comes that major network operators 'cannot' get additional
>address-space assignments -they- need, because of a lack of 'unassigned'
>address-space, *AND* there are significant blocks of 'unannounced' space,
>one *will* see operators starting to use that space, regardless of what
>the 'authorities' decree.

There's a number of interesting options at that point, and ISP's
*will* do what's necessary to keep their businesses running. 
The challenging part is whether the enlightened self-interest
will result in an actual functional result for the Internet.

For example, one option would be for ISP's and deep-pocketed
new endeavors to start mining the unannounced address space. 
This has some fairly interesting side-effects, as there's every
reason for "holders" who have no other use for their space to
heavily subdivide their blocks for maximum financial return.

Since there is no inherent hierarchy to the space obtained in
this manner (whether by ISP or end-sites), we need to expect
a much higher ratio of new routes to new Internet customers.
Obviously, if "enlightenment" exceeds "self-interest", the ISP
community would also need to get together and come up with
some guidelines for avoiding the routing table explosion, but
but that's always been a hard meeting to convene (and we've
no mechanism enforce the outcome).

I fully agree with you; The ISP's & Internet community get to
call the shots here, just as they get to set the Internet resource
policies which are followed by the RIR's.  I wouldn't be surprised
to see the ISP community decide that it's very much in their
interest to know the utilization of the unannounced blocks,
whether it's so that they can later play "let's make a deal" or so
that they can direct the current IANA to invalidate per RFC2050
and get the space through today's processes.

>A coalition of network (and IX) operators could decide _tomorrow_ to ignore
>*all* address-range "assignments" from the above-mentioned hierarchy, and
>only route traffic from address-blocks "blessed" by some alternative source,
>and there is *nothing* that the aforementioned 'authorities' could do to
>prevent it.

Perfectly true, although it certainly would be very, very exciting for
the "alternative source", who would be relying entirely on some
expression of collective will of the ISPs when reclaiming space... 
The time from appropriation to courtroom might be really quick,
unless there was also an accepted community statement (formed
in an open & public fashion) which provided the basis for action.

/John