Re: [v6ops] [EXTERNAL] Re: Scope of Unique Local IPv6 Unicast Addresses (Fwd: New Version Notification for draft-gont-6man-ipv6-ula-scope-00.txt)

[Ted Lemon <[e-mail]>]

Fernando, I’m not going to retread this argument again. It’s clear that you and I disagree about what “global” means. You seem to want it to be some kind of guarantee, but it’s not. What’s also clear is that there isn’t any kind of consensus that there’s a problem here of the magnitude that you are suggesting. Your sole basis for saying that there’s a problem is that the definition of “global” in RFC 4007 doesn’t match the use of “global” in RFC 4193. I agree with you that this is the case, but this is not a problem statement that requires standards action. If someone is confused, you already know how to explain it to them.

I’m not going to participate in this discussion any further unless it becomes clear that the working group wants to take some action that will make things worse, in which case I will participate in the sense of opposing such action.

> On Feb 15, 2021, at 12:03 AM, Fernando Gont <[e-mail]> wrote:
> 
> On 14/2/21 19:25, Manfredi (US), Albert E wrote:
>> -----Original Message-----
>> From: ipv6 <[e-mail]> On Behalf Of Fernando Gont
>>> ULAs can't be global.
>> I think we've been going around and around on this.
>> Maybe it's good to compare the ULA situation with RFC 1918 private IPv4 addresses. In such a comparison, it is clear that RFC 1918 addresses "can't be global," and must be used only within some admin domain. But it is also clear that the intention, or hope, of ULAs, is that they be globally unique.
> 
> The intention is that, *given a subset of ULAs* the probability of collision is low.
> 
> 
> 
>> Yes, in practice, we know that there is a non-zero probability that ULAs won’t be globally unique, but at the same time, it is false to claim that ULAs "can’t be" globally unique. 
> 
> Please compute the birthday paradox assuming that each CPE in the world locally-generates a ULA prefix, and share the math with us. (that's what global scope means as per RFC4007)
> 
> Spoiler: P~1
> 
> 
>> No one is telling us, "Go ahead and duplicate your ULAs, among admin domains, because they will be filtered out at border routers anyway."
>> At most, I'd add somewhere a notion of "only guaranteed within an administrative domain," and be done with it. I just don’t see how that is so confusing for anyone.
> 
> As per RFC4007:
> 
> scope means: topological span where the address is unique.
> global means: Internet-wide span
> 
> 
> Then we have this notion of "ULAs are 'global scope'", in which "global" meaning something different than what "global scope" means as per RFC4007. -- Hence the incongruence.
> 
> e.g., if "global" means "an administrative domain", then that's not global -- unless you're also keen to argue that link-locals can be considered "global scope".
> 
> Thanks,
> -- 
> Fernando Gont
> SI6 Networks
> e-mail: [e-mail]
> PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492
> 
> 
> 
> 
> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops