Re: [address-policy-wg] 2011-05 Discussion Period extended until 14 February 2012 (Safeguarding future IXPs with IPv4 space)
I have a direct interest in this proposal. Having got that out of the way, I generally agree with principal of the proposal. If a bunch of organisations want to bunch together and set up some common infrastructure for interconnection, this is probably more important than giving public ip to a tiny number of dial up end users. One suggestion:
Assignments will only be made to IXPs who have already applied for, or received an IPv6 assignment for their peering LAN.
This is pointless. You can't make people use IPv6 and politically stuffing it down the throat of an IXP won't do anything other than cause facepalming. It certainly won't cause ipv6 to be more rapidly deployed. Either IPv6 will happen because of necessity or it won't. If IXP participants want ipv6, then they will push the exchange operator to implement it, and it will be done. But mandating an ipv6 assignment request in an ipv4 assignment policy is, well, silly. Also, a nit:
This space will be used to run an Internet Exchange Point peering LAN, other uses are forbidden.
The comma should be replaced with either a semicolon or else a full stop. Nick
Hi! On 01/17/2012 10:54 PM, Nick Hilliard wrote:
I have a direct interest in this proposal. Having got that out of the way, I generally agree with principal of the proposal. If a bunch of organisations want to bunch together and set up some common infrastructure for interconnection, this is probably more important than giving public ip to a tiny number of dial up end users.
Sorry but no, that's not acceptable. ISPs or IXPs are users of IP addresses. Just like anyone else. Actually ISPs have a bunch of ways to deal with situations where they can't get as much address space as they'd wish to. The generic enduser doesn't. A fair distribution had to be based on equal distribution. How to define this is ofc arguable - but if there's an enduser-need (or, say, 16 of them) e.g. for a /28, which is just rejected while a hand full of ISPs (together already using maybe several /16) get further /24s - that'd just be outrageous. Nothing against you personally - but i think the fact that this proposal has quite a crowd of supporters here, of which actually virtually all are IXP-persons, while the majority of non-IXP-persons is simply hardly represented at all, is a severe legitimacy problem. I think doing 'net community' work that way is not a good thing - that way lies suffering.
This is pointless. You can't make people use IPv6 and politically stuffing it down the throat of an IXP won't do anything other than cause facepalming. It certainly won't cause ipv6 to be more rapidly deployed.
Either IPv6 will happen because of necessity or it won't. If IXP participants want ipv6, then they will push the exchange operator to implement it, and it will be done. But mandating an ipv6 assignment request in an ipv4 assignment policy is, well, silly.
I agree on that one. While we're at it, on proposals in general: That "Rationale"-section of course always has an "a) Arguments supporting the proposal" paragraph, where the authors try to sell their standpoint. Usually - this case is no exception - at least in part with blatantly wrong information. The "b) Arguments opposing the proposal" paragraph practically always consists of: "None." (which is probably almost never true) - experience shows that this paragraph also never changes, not if arguments pop up in the discussion of the proposal, not even if opposers point to this paragraph and suggest the respective arguments should be listed there. That's pathetic. If proposers regularly don't have the balls to also mention the opposing arguments, then just drop the "Rationale"-section, for dignity's sake... Regards, Chris
On 18/01/2012 17:32, chrish@consol.net wrote:
A fair distribution had to be based on equal distribution. How to define this is ofc arguable - but if there's an enduser-need (or, say, 16 of them) e.g. for a /28, which is just rejected while a hand full of ISPs (together already using maybe several /16) get further /24s - that'd just be outrageous.
Hi Chris, I'd be very interested to hear more on your ideas about how to define fairness. I've tried myself on a number of a occasions to come up with some clear ideas about what constitutes "fair" in an environment of scarcity, but unfortunately have made very little progress. Certainly I'd agree that it's very easy to define what's unfair. That would be when someone received more IP address space than they ought to have got. But fairness? Not so easy. Anyway, you obviously have some pretty clear ideas about all of this, so would you be able to share them? The second last /8 will run out in the next couple of months, and we're going to have the most unholy bunfight when it does. Obviously it would be much better to come up with a final definition of fair while people are still quite mellow about v4 resource allocation. Nick
Hi! On 01/18/2012 07:15 PM, Nick Hilliard wrote:
I'd be very interested to hear more on your ideas about how to define fairness. I've tried myself on a number of a occasions to come up with some clear ideas about what constitutes "fair" in an environment of scarcity, but unfortunately have made very little progress.
I believe defining fairness is undoubtedly always a challenging issue. But as i already wrote: dividing into groups and handling them unequally is _definitely_ an increase in unfairness. So what i can define in the issue at hand is: It is more fair not to introduce privileges for such groups than otherwise. So let's just not make IXPs/ISPs more equal than others.
would you be able to share them? The second last /8 will run out in the next couple of months, and we're going to have the most unholy bunfight when it does. Obviously it would be much better to come up with a final
Humm - we're talking about a proposal to make a small group special in the mentioned case: We are already in the unholy bunfight (cheers!). But you asked about the "IPv4 depletion" issue. As you seem interested, as a little hint, just take a look e.g. into the first /2 ... And there are a lot more issues of that kind... Regards, Chris
On 19/01/2012 12:56, Chris wrote:
But as i already wrote: dividing into groups and handling them unequally is _definitely_ an increase in unfairness.
So, one person or company, one IP address? Nick
Hi! On 01/19/2012 02:11 PM, Nick Hilliard wrote:
But as i already wrote: dividing into groups and handling them unequally is _definitely_ an increase in unfairness.
So, one person or company, one IP address?
If that's what you want to give to anybody, that should be what you give to IXPs/ISPs. (You probably won't find anybody else supporting this...) Regards, Chris
On 19/01/2012 12:56, Chris wrote:
But as i already wrote: dividing into groups and handling them unequally is _definitely_ an increase in unfairness.
So, one person or company, one IP address? Nick
participants (4)
-
Chris -
chrish＠consol.net
-
Nick Hilliard -
Nick Hilliard