[arin-ppml] GUA vs ULA vs ?

[Steve Bertrand]

On 2010.03.30 16:53, David Farmer wrote:
> michael.dillon at bt.com wrote:

>>> Personally, I'd like to have the IETF define FC00::/8 for this
>>> purpose, and delegate it to IANA to allocate to the RIRs 
>>
>> RFC 4193 is very muddy when it comes to laying out how the
>> FC00::/7 addresss are defined. First of all, the FC00 part
>> is hexadecimal, so each letter refers to 4 bits of the address. Given
>> the /7 part, this means that only the F and
>> 3 bits from the C, are fixed. The L bit, which determines
>> whether it is ULA-C or ULA-RANDOM (called ULA-L by IETF folk)
>> is part of that C digit. This means that addresses beginning
>> with FD, are ULA-L (ULA-RANDOM). Addresses beginning with
>> FC are currently in limbo because they are defined by the
>> RFC, but not yet assigned to IANA.
>>
>> So, FD00::/8 are the ULA-L randomly assigned unique unicast
>> addresses intended for local use.
>>
>> And FC00::/8 are indeed the ULA-C addresses that need another
>> RFC in order to instruct IANA to allocate them to RIRs.

My understanding from what I've been reading today is that ULA-C has a
split audience that goes well beyond the scope of our ARIN RIR
discussion. It seems as though there is no consensus on this matter on
any level. I hope that my perception is correct regarding this issue.

>> Sadly, no. Nor are the old 6bone addresses in their bogon list.
>> Not even SIXXS <http://www.sixxs.net/tools/grh/bogons/> includes
>> these as bogons.
> 
> I'll admit I didn't look that closely, but when I Googled it (Cymru
> ipv6) I got the following;
> 
> http://www.cymru.com/Bogons/ipv6.txt
> 
> FC00::/7 is listed under 3-1-1-1-2 for ingress packet filter and
> 3-1-2-1-2 ingress prefix filter.  You are right that they don't have the
> old 6Bone in the filter though.

Does this not cover it? (search for 3000::/4):

http://www.cymru.com/Documents/fullbogon-ipv6.txt

Also, I am extremely confident that it won't be long before these
addresses appear in Team Cymru BGP IPv6 BOGON feeds. From what I can
tell, 3000::/4 covers the 6Bone space (x:x:1::ffff is my null route),
hence, I learn this:

rtbh-trig2#sh ipv6 route 3ffe::1

IPv6 Routing Table - 20696 entries
B   3000::/4 [20/0]
     via 2607:F118:1::FFFF

>>> Your /32 presumably is a GUA-PA provider allocation not an GUA-PI
>>> end-user assignment. 
>>
>> Do we really need to change PA and PI to have a GUA prefix?
> 
> No, write me up a definitive style sheet for PPML post and I'll follow
> it. :)

Is there at least a consensus on _why_ there is no consensus? Does it
boil down to the potential abuse of connecting a non-connected block?

Steve