RFC 4291, "IP Version 6 Addressing Architecture", February 2006

Note: This RFC has been updated by RFC 5952, RFC 6052, RFC 7136, RFC 7346, RFC 7371, RFC 8064

Errata ID: 4406
Status: Rejected
Type: Technical
Publication Format(s) : TEXT
Reported By: Peter Paluch
Date Reported: 2015-06-29
Rejected by: Brian Haberman
Date Rejected: 2015-06-30

Section 2.5.6 says:

   Link-Local addresses are for use on a single link.  Link-Local
   addresses have the following format:

   |   10     |
   |  bits    |         54 bits         |          64 bits           |
   +----------+-------------------------+----------------------------+
   |1111111010|           0             |       interface ID         |
   +----------+-------------------------+----------------------------+

It should say:

   Link-Local addresses are for use on a single link.  Link-Local
   addresses have the following format:

   |   10     |
   |  bits    |         54 bits         |          64 bits           |
   +----------+-------------------------+----------------------------+
   |1111111010|     arbitrary value     |       interface ID         |
   +----------+-------------------------+----------------------------+

Notes:

Section 2.4 of this RFC states that link-local addresses are identified by having the prefix FE80::/10. According to this prefix, the entire range of link-local addresses covers all addresses within the range of FE80:: through FEBF:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF.

The format of link-local addresses as stated in Section 2.5.6 seems to partially contradict Section 2.4 in that it fixes the internal 54 bits between the 10-bit prefix and the IID to the value of 0. If that was the case then the prefix of link-local addresses would have been in fact FE80::/64 instead of FE80::/10, and the range of link-local addresses would have been limited to FE80:: through FE80::FFFF:FFFF:FFFF:FFFF. Thus, the definition of link-local addresses as follows from Section 2.4 does not align with the definition of link-local addresses as presented in Section 2.5.6.

I suggest resolving this internal contradiction by explicitly stating in Section 2.5.6 that the internal 54 bits of a link-local address that follow the FE80::/10 prefix can be of an arbitrary value.

Thank you!
--VERIFIER NOTES--
Section 2.4 reserves the FE80::/10 prefix for use on the local link. Section 2.5.6 specifies that the current definition of a link-local address falls in the FE80::/64 prefix. The suggested change would be a technical change to a consensus-based decision. The re-definition of the link-local address would need to be driven by a draft updating RFC 4291.

Report New Errata