Hello Tatuya,
Thank you for the careful review. Follow-up below:
On 1/6/2017 11:08 AM, Tatuya Jinmei wrote:
- Section 4.1: I guess the MNID is generally supposed to be unique
(at
least in the realm the ID is used), but not all IPv6 addresses are
guaranteed to be unique (a link-local or unspecified address is an
obvious example, an ULA may also be inappropriate depending on the
usage context). It may be better to note the fact, and you may
also
want to impose some restrictions on the type of address that can be
used as an MNID.
This is correct. I will fashion some language as suggested. I think it
is appropriate to allow ULAs, but multicast and unspecified addresses
seem clearly inappropriate, and I am i favor of disallowing link-local
addresses.
- Section 4.5
2000, modulo 2^32. Since the link-layer address can be of
variable
length [RFC2464], the DUID-LLT is of variable length.
I don't understand why RFC2464 is referenced in this context. This
RFC is about IPv6 over Ethernet, and assumes a fixed (6 bytes)
length of hardware address.
I don't quite know what to do about this. I actually just copied this
language from RFC 3315. I think that the citation is also wrong in RFC
3315, for the same reason as given here. I could simply delete the
reference to RFC 2464.
- Section 4.9: s/is (GRAI)/(GRAI)/
The Global Returnable Asset Identifier is (GRAI) is defined by the
Fixed. I also checked for other similar instances and did not find any.
Regards,
Charlie P.