Re: [IANAxfer@apnic] Key elements of the transition ofIANA stewardship
Please see below.
Thanks and best,
Richard
> -----Original Message-----
> From: ianaxfer-bounces@apnic.net [mailto:ianaxfer-bounces@apnic.net]On
> Behalf Of John Curran
> Sent: vendredi, 12. septembre 2014 06:12
> To: Avri Doria
> Cc: ianaxfer@apnic.net
> Subject: Re: [IANAxfer@apnic] Key elements of the transition ofIANA
> stewardship
>
>
> On Sep 11, 2014, at 11:09 PM, Avri Doria <avri@acm.org> wrote:
>
> > On 11-Sep-14 08:33, John Curran wrote:
> >> Again, feel free to run one if you'd like, and assign
> >> the entire 32-bit IP address space to various parties in
> "Richard's Global
> >> IP Registry"
> >
> > Intriguing question.
> >
> > And a little mean.
>
> Apologies - I was trying to be factual in nature, as Richard
> postulated some
> top down "authority" as dictating how IP addresses must be used.
I apologize for not being sufficiently clear. I did not mean to postulate a
top-down authority dictacting how IP address must be used. Indeed, I'm well
aware of the fact that there is no such authority (or at least I don't know
of any).
What I meant is that, within the current voluntary system, there are bodies
that make decisions. Those are the RIRs. Each RIR has its own process. For
example, in ARIN the decisions are ultimately made by the ARIN Board, unless
I am mistaken.
At the highest level, decisions are made by the ICANN Board. At present,
the IANA functions contract compels the ICANN Board to adopt the RIR
policies unchanged. Absent that contract, it seems to me that ICANN would
be free to change the RIR policies.
Implementation of numbers allocated under the current system (or any future
system) is of course a voluntary matter.
> In truth, it
> is the other way around, wherein the Internet number Registry
> system provides
> a functional service of globally coordinated uniqueness, so
> Internet service
> and content providers find it useful to make use of it. There is
> no ability
> by anyone to command its use, but the availability of widely-used
> globally
> unique addresses is rather compelling.
Correct.
And I agree with what you write below.
>
> > Nonetheless, a thought experiment since you seem to posit something as
> > within the realm of possibility.
> >
> > What would it take for Richard and Friends to acquire control over some
> > IP address space; Either v4 (maybe they get it from the legacies) or v6
> > (maybe they just ask for an unassigned block). Would they be allowed to
> > allocate the v4 (perhaps one way for MENA and other v4 deprived
> > communities to get it)? Would they be allocated by IANA a block of v6?
> > If not, why not? Especially since there are enough bits for an infinity
> > of universes.
>
> Avri - You probably want to read RFC 7020 for background on the Internet
> Number Registry System. While everyone is familiar with IANA and
> the RIRs,
> it is also true that there is an entire system of ISPs and local Internet
> registries (LIRs) that perform assignments in turn to their customers
> (usually, but not always, as part of their Internet service bundle.)
>
> If your address block is unique with respect to the everyone
> else's in the
> RIRs and IANA registries, then it's part of the Internet Number Registry
> System. If you have addresses that are only unique with respect
> to a small
> number of others because _you_ are maintaining an independent
> directory of
> who can uniquely use each possible 32-bit IPv4 address, then you have a
> created a completely independent registry.
>
> It would be fairly difficult to start issuing IPv4 space from the Internet
> Number Registry System, because the IANA is effectively in the final phase
> of IPv4 and has no real available resources to issue to any
> Internet number
> registry (RIR or otherwise.)
>
> > And what if people around the world, perhaps members of global
> > organizations or just a bottom-up groundswell of contrarians, applauded
> > and said yes, we want IP allocations from Richard and Friends? Would
> > that be allowed? Who needs to allow it? IANA? How do they decide? The
> > RIRs? why? Because they were there first?
>
> Actually, there have been organizations that have run completely distinct
> IPv4 address registries (effectively keeping their own list of
> who gets to
> use each possible unique address, without regard to the IANA and
> the RIRs);
> this was what I meant when I pointed out to Richard that it is indeed
> possible to do, but you only have effective communication with community
> that decides to follow that same registry (since there are conflicts when
> trying to interconnect with those using the Internet Number
> Registry System.)
> In the past, this sort of approach of private registry was done in order
> to provide a very large address space for supplier networks to
> operate, i.e.
> effectively making a very large virtual private network. I don't
> think these
> are in operation now, as the overhead of keeping some of their
> machines on
> 'real' Internet addressing and some on 'privately coordinated' Internet
> addressing is significant effort.
>
> But it does work, anyone can use the Internet protocols with any IP
> addresses they like, it's just advised that you coordinate with those
> who wish to interconnect with regarding choice of numbers. The vast
> majority of ISP's and their customers find making use of the Internet
> Number Registry System a convenient way to do that coordination...
>
> > Would that be allowed? Who needs to allow it? IANA? How do
> they decide? The
> > RIRs? why? Because they were there first?
>
> Of course it's allowed, how could it not be? You can use IP protocols
> with any numbers you wish, but your realm of interoperability extends
> only as far as your scope of your Internet addressing coordination.
>
> Just as you use any address block you wish in your private networking
> lab, you can do the same for your corporate network or among your and
> your associates; just remember, you'll only be able to have reliable
> communicate with others who coordinate via the same Internet address
> registry, whether that's the "default" one provided for in RFC 7020
> or a completely separate coordination of your own devising.
>
> > If the people wanting it are enough for authority - and I agree that is
> > the way it should be, then how could Richard and Friends make
> this happen?
>
> Yes, completely independent registries are possible, with a usefulness
> which matches the scope of coordination. No one can prevent you from
> configuring your equipment with any IP address in the world, you just
> have to realize that if it is not coordinated globally, the addresses
> are not going to be useful for communicating with others globally.
>
> With respect to new registries within the existing Internet Number
> Registry System, discussions about such registries take place in the
> RIR and ICANN communities via development of global number resource
> policies. It is not just uniqueness that is needed to make an address
> block useful, but it is also having the ISP community decide that they
> will route address blocks as they are issued within the registry system.
> As such, discussions of how registries come into being, how IANA issues
> address space to them, and how they issue in turn to service providers
> and end-users is of interest to the entire community and have to be
> global policy since the routing implications are global in nature.
>
> FYI,
> /John
>
> John Curran
> President and CEO
> ARIN
>
>
>
> _______________________________________________
> IANAxfer mailing list
> IANAxfer@apnic.net
> http://mailman.apnic.net/mailman/listinfo/ianaxfer
>