> On 20 Oct 2020, at 19:30, ripedenis--- via address-policy-wg
> <[email protected]> wrote:
>
> Legacy space is a pain and should be normalised at every opportunity. Because
> of the market this has become a huge financial asset. If the holders want to
> cash it in, once it is sold it should lose this special, untouchable status.
Why? What’s the rationale for that? What good will it achieve? What problem
does that fix?
If your concern is it’s too much of a resource drain on an RIR to track these
transfers, then let’s first quantify the problem before deciding how to solve
it. My gut feel is those overheads are marginal and also a tolerable part of
doing business. It may well be less hassle to just let those costs be lost in
the noise than trying to invent some sort of cost recovery scheme. Which would
of course provide lots of opportunities for shed-painting and rat-holing.
> The transfer policies are their means to sell it. These policies should
> insist on sold legacy space being normalised and subject to all RIR policies.
Denis, I *strongly* disagree. Legacy space is like an RS-232 lead*: an
annoyance from a bygone era that will always be with us. :-) We just have to
suck it up. Unless of course one day the Internet stops using IPv4 and
everyone’s on IPv6. [Who’s sniggering at the back!?!]
* Kids, ask your grandparents...
What’s the justification for "normalised at every opportunity” and what do you
mean by that anyway?
Forcing transferred legacy space to be subject to RIR policies is utterly wrong.
First, the space was issued before the RIR system existed. It shouldn’t be
subject to what amounts to retrospective legislation. That probably wouldn’t
survive the flimsiest of challenges in the courts. Besides, we agreed that
principle during the ERX exercise some years ago when European legacy space
moved from ARIN to the NCC. Those legacy holders were not made to pay NCC fees
or had their holdings of legacy space influence how their future IPv4
allocation/assignment requests got handled. Why go back on this principle now?
What’s changed since then to justify that?
Next, if transfers involving legacy space were forced to be subject to RIR
policies, you’d just drive those transfers underground. Or the parties involved
would contrive “mergers” and “reorganisations” to conceal the truth that
addresses changed hands. That would be bad in far too many ways to list here.
What’s more important, maintaining an accurate database of address space
holders or upholding the purity of some doctrine about nearly dead IPv4 address
policies that are irrelevant to a post-v4 world?
FWIW we abandoned that notion of ideological purity when the LIR transfer
policies were introduced. The consensus then (and now) was maintaining accurate
info about who held address resources was more important than following a no
longer credible policy of forcing LIRs to return their surplus space to the NCC
for redistribution.
Finally, suppose the recipient of a legacy transfer is not an LIR. Your
suggestion implies they’d have to become one. That’ll attract the attention of
legislators, regulators and the competition authorities faster than you can say
anti-trust lawsuit.