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

[Fernando Gont <[e-mail]>]

On 13/2/21 01:51, Ted Lemon wrote:
> On Feb 12, 2021, at 10:13 PM, Fernando Gont <[e-mail] 
> <mailto:[e-mail]>> wrote:
>> You might argue that architecture is not important (I'd disagree with 
>> that :-) ), but, at the end of the day, if it's not important, why 
>> pretend to have one in the first place? (in particular when there's no 
>> consistency with the actual protocols)
>>
>>
>> As noted, the only practical implications I've seen have been:
>>
>> * Folks willing to have a registry for ULAs, on the expectations that
>>  they are indeed unique -- whcih they are not!
>>
>> * Folks having a hard time understanding the addressing architecture --
>>  in particular the scoped addressing architecture (RFC4007) and how
>>  that applies to actual specs (e.g., definition of ULAs)
> 
> The thing is, the meaning of the architecture document as it pertains to 
> ULAs is plain to me, and it’s just what Brian said.

If by "plain" you mean "clear", then this discussion should be an 
indication that it's not. We have a definition of scope in RFC4007 which 
doesn't match RFC4193... and then it should be clear that we all have a 
different understanding of what "scope" and "global scope" (in 
particular) mean.



> So who is confused? 

Well, I certainly have my own take on the topic. The fact that most of 
the folks that have participated in this thread have their own and 
different take, probably means something.



> Are you having operational issues, or just speculating? 

There are certainly no operational issues that I know of. (But we also 
don't really leverage IPv6 addressing anyway, so...)

As noted, I believe that the confusing terminology does have concrete 
implications. e.g. see:

---- cut here ----
5.1.  Address Attributes in Programming Languages

    Python's ipaddress library [Python-ipaddr] defines 'IPv6Address'
    objects that have a number of attributes, including:

    o  'True' if the address is allocated for private networks.

    o  'True' if the address is allocated for public networks.

    For ULAs, the is_private attribute is 'True', while the is_global
    attribute is 'False'.  This contradicts the definition of ULAs as
    having "global scope" [RFC4291] [RFC4193], but is in line with the
    specification update performed by this document (see Section 6).
---- cut here ----

(this one is documented in our draft, but credit goes to Brian for 
spotting it).

Are the is_private and is_global attributes employed by such library 
correct? Do they match our specs? Do they match the obvious 
interpretation of such attributes?  Should they be changed? Why, or why not?



> I don’t mean to minimize your concern—just trying to get a 
> sense of why this has come up. Do you not know how to operate in the 
> presence of ULAs? Have you been getting questions from customers that 
> you can’t answer to their satisfaction? 

I have got questions from customers that I wasn't able to answer with 
*my* own satisfaction -- let alone theirs, so to speak. :-)

Explain the scoped addressing architecture (even the very definition of 
"scope" means.) Then introduce ULAs, and note that they are global 
scope. Both cannot be right.

Somewhat related:

in the context of draft-ietf-6man-slaac-renum, some argued that one 
should differentiate between GUAs and ULAs. And that you should only 
phase out stale ULAs if you have fresh ULAs. And that you should only 
phase out stale GUAs if you know of fresher GUAs.

The rationale of this has a lot to do with whether GUAs and ULAs are a 
different sort of animal.

Thanks,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492