Re: [Technical Errata Reported] RFC5453 (5589)

[Brian E Carpenter <[e-mail]>]

On 2019-01-03 13:54, Fernando Gont wrote:
> On 2/1/19 18:51, RFC Errata System wrote:
> 
> [....]
>> Corrected Text
>> --------------
>>    +-----------------------------------------+-------------------------+
>>    |        Interface Identifier Range       |       Description       |
>>    +-----------------------------------------+-------------------------+
>>    |           0000:0000:0000:0000           |  Subnet-Router Anycast  |
>>    |                                         |        [RFC4291]        |
>>    |                                         |                         |
>>    | FDFF:FFFF:FFFF:FF80-FDFF:FFFF:FFFF:FFFF | Reserved Subnet Anycast |
>>    | FFFF:FFFF:FFFF:FF80-FFFF:FFFF:FFFF:FFFF |    Addresses[RFC2526]   |
>>    +-----------------------------------------+-------------------------+
>>
>>                        Table 1: Current Assignments
> 
> How should we make this into the actual IANA registry?

Well, first some higher authority would have to approve the erratum.
Then I'm pretty sure the AD could simply ask IANA to make the update.

Another interesting question is who will instruct implementers
of RFC7217 to add a check for this case.

>> Notes
>> -----
>> Both these ranges are in fact reserved by Section 2 of RFC2526. Under current recommendations [RFC7217] the FDFF range is no longer recommended for routine use, and the FFFF range is equally likely to occur in a pseudo-random 64-bit IID. Credit to Kerry Lynn for pointing this out.
> 
> The FDFF range still seems to be recommended for routine use of its
> intended usage (i.e., anycast) -- that's why we are avoiding it in e.g.
> RFC721> 
> OTOH, the FFFF range wasn't present in the IANA registry, and might be
> being employed for pseudo-random-IIDs.

Yes, but it's a pretty unlikely problem. I think if we fix the text
and IANA registry now, it's really not going to affect the real world
much.

Rgds
   Brian

> 
> Thoughts?
> 
> Thanks,
>