[arin-ppml] Draft Policy ARIN-2019-18: LIR/ISP Re-Assignment to Non-Connected Networks - Clarifying Language

Owen DeLong owen at delong.com
Mon Nov 4 10:31:20 EST 2019



> On Nov 4, 2019, at 07:11 , scott <scott at solarnetone.org> wrote:
> 
> Hi Owen,
> 
>>> pardon, we are talking about leasing to someone not operating a network, hence the "non-connected systems in the draft title".
>>> nobody has a problem with upstream provided addresses via a standard dhcp "lease”.
>> 
>> I don’t believe that’s true. We are talking about leasing to someone whose network is not connected to the network of the lessor.
>> 
>> That’s not the same as someone who doesn’t have a network.
> 
> Fair enough.  Similar, but not identical to "number portability" in the PSTN then.

I guess, kind of. I think it’s a rather tortured analogy.

>>> How is this cheaper than addresses provided by upstream?  Granted, it can be costly to roll your own routing infrastructure on addresses allocated to you from the RIR, particularly if you don't have the technical chops to do it yourself.  That said, you are going to have to announce your "leased" address space somehow anyway.
>> 
>> One example I can think of is a situation where you don’t want to buy connectivity from $LARGE_PROVIDER, but $SMALL_PROVIDER doesn’t have addresses for you to lease.
> 
> This assumes $SMALL_PROVIDER is willing to announce your leased space, and you are leasing a /24 or greater, so it is routable under current practiced designed to keep the size of the routing table managable. If you are announcing yourself, then this need is fulfilled by existing methods in the NRPM.

Not entirely, IMHO.

$SMALL_PROVIDER is likely to be very willing to announce (or accept your announcement of)  your leased space if you can show an indication that it is legitimately leased to you and you are paying them for connectivity. The bigger question would be the willingness of $LARGE_PROVIDER to lease their space rather than try to use it to force connectivity purchases, but this question delves into legal areas where I am _NOT_ an expert, such as the Sherman act. The point is that, assuming there are parties
that have space and are willing to lease it to customers not purchasing connectivity (and there definitely are), there are reasons for doing so that are consistent with the general spirit of ARIN policy and that the requirement to associate such a lease with the
provision of connectivity services is not necessarily ideal in every circumstance.

>> In such a case, it may be desirable to lease the addresses from $LARGE_PROVIDER while purchasing the attendant connectivity from $SMALL_PROVIDER.
>> 
>> One of the complaints we heard in rather strong terms from several members of the ARIN community recently in Austin was that $SMALL_PROVIDERs are at a disadvantage compared to $LARGE_PROVIDERS because they were not able to obtain as much extra address space prior to runout.
> 
> While the early bird gets the worm, the late bird should not be forced to starve.  I get it.  The late bird, however, does have the option to visit wormfarm v2.0, where worms are plentiful, and standard policy is all you can eat for anyone at the table.

Yes, but when you’re not eating the worms, but instead have to use them to connect to other birds, including early and late ones, and some of the early birds won’t talk to you without worms from worm farm v1.0, the situation is a bit different.

I doubt you can find anyone more interested in seeing IPv4 retired and IPv6 as the dominant lingua franca of the internet than me. I’ve spent the last 10 years of my life working in various ways to push the deployment of IPv6.

Nonetheless, the reality today is you still can’t get to the Amazon store without IPv4. You still can’t Skype without IPv4. You still can’t contact most banks’ web sites without IPv4…

Like it or not, wormfarm2.0 is an important part of a complete connectivity solution, but you still need some wormfarm 1.0 products if you don’t wish to be isolated to a subset of the internet.

> 
>> 
>> This wouldn’t entirely level the playing field for this scenario, but it might provide at least some relief.
> 
> I think we can all agree that a level playing field is an honorable and ultimately attainable goal.

Absolutely.

Owen

> 
> Thanks,
> Scott
> 
> 
>> 
>> Owen
>> 
>>> 
>>>> Cost wise, its effective. While I agree
>>>> the business model may be less desired to some, the outcome is legit.
>>>> The question could be about accurate tallying of utilization.
>>>> Best,
>>>> -M<
>>>> 
>>>> On Sun, Nov 3, 2019 at 17:58 scott <scott at solarnetone.org> wrote:
>>>>     IMHO, we should do everything we can to prevent "internet
>>>>     landlords."
>>>>     Further, I do not see a legitimage use case problem that is
>>>>     solved by
>>>>     allowing leasing that is not solved by upstream provided
>>>>     address space, or
>>>>     barring that, 4.10 of the NRPM.  If we want to enable spammers,
>>>>     attack
>>>>     networks, and other bad actors, then leasing is for sure a
>>>>     great solution
>>>>     for them, and the "internet slumlords" that would provide their
>>>>     resources.
>>>> 
>>>>     Scott
>>>> 
>>>>     On Sun, 3 Nov 2019, Martin Hannigan wrote:
>>>> 
>>>>    >
>>>>    >
>>>>    > On Sat, Nov 2, 2019 at 10:30 PM Owen DeLong <owen at delong.com>
>>>>     wrote:
>>>>    >
>>>>    >
>>>>    > [ clip ]
>>>>    >
>>>>    >       However, what I do not want to see is a situation where
>>>>     we
>>>>    >       permit the desire to lease space as a justification for
>>>>    >       obtaining space through the transfer market (or
>>>>    > any other mechanism). If you want to leas space you already
>>>>     have,
>>>>    > then fine. But the desire to lease space in and of itself
>>>>     should not
>>>>    > qualify as “utilization” or
>>>>    > “need” in evaluation of any form of resource request.
>>>>    >
>>>>    >
>>>>    >
>>>>    > Needs a little more clarify for me. Either the lessor or
>>>>     lessee has a right
>>>>    > to use the numbers as justification? The lessee may be the
>>>>     logical party,
>>>>    > but seems less likely to be in the transfer market. However,
>>>>     if they are
>>>>    > leasing numbers they may have legitimate need. On the other
>>>>     hand, if a
>>>>    > lessor has a ratio like an ISP or other provider using
>>>>     numbers in an
>>>>    > aggregated manner _and_ the lessee can't use the lease as
>>>>     justification for
>>>>    > transfers, that would seem to be inline with current
>>>>     practice. I do think
>>>>    > legitimately "in use" addresses should be eligible for "need"
>>>>     credit. Isn't
>>>>    > the idea that "access" is being facilitated by providing the
>>>>     numbers? You
>>>>    > can use RFC 1918 address space as a justification for need
>>>>     and the numbers
>>>>    > are technically "not connected". I'm thinking source nor
>>>>     business model
>>>>    > should matter, but that we're careful who is getting credit
>>>>     for them. Just
>>>>    > saying that made me wonder if this is even worth addressing.
>>>>    >
>>>>    > Feels like it is more sensible to allow the both to
>>>>     demonstrate use as a
>>>>    > justification and let ARIN process sort it out.
>>>>    >
>>>>    > $0.02
>>>>    >
>>>>    > Best,
>>>>    >
>>>>    > -M<
>>>>    >
>>>>    >
>>>>    >
>>>>    >
>>>>    >
>>>>    >
>>> _______________________________________________
>>> ARIN-PPML
>>> You are receiving this message because you are subscribed to
>>> the ARIN Public Policy Mailing List (ARIN-PPML at arin.net).
>>> Unsubscribe or manage your mailing list subscription at:
>>> https://lists.arin.net/mailman/listinfo/arin-ppml
>>> Please contact info at arin.net if you experience any issues.
>> 




More information about the ARIN-PPML mailing list