On Fri, 24 Aug 2001, Bob Hinden wrote:
> Please send substantive comments to the ipng mailing list, and minor
> editorial comments to the authors. This last call period will end one
> week from today on August 31, 2001.
2.1 Addressing Model reads:
A
single interface may also be assigned multiple IPv6 addresses of any
type (unicast, anycast, and multicast) or scope.
I don't think it's good practise to talk about assigning multicast
addresses to an interface. This may confuse people, and make people think
of doing stuff like 'ifconfig xl0 inet6 ff0e::1 alias' ie. actually assign
it to as a regular IPv6 address, which might confuse source address
selection a bit and be a little funky to say at least.
Editorial nit: first CIDR reference is in 2.3.
2.5.5 IPv6 Addresses with Embedded IPv4 Addresses:
This discussion, dating back a few years, is based on times when there
were only compatible/mapped addresses which embed IPv4 addresses. Now
there are more, and more on the way. I'm not sure whether it's sensible
to include all of them in this draft, but a decision should be made
whether the current ones are really that much needed either or enough,
given others also exist.
Personally I think mapped address, stuff like SIIT aside, should be
fundamentally different than compatible address, and at least mapped addr
should be discussed in this draft. Compatible is just a transitionary
mechanism that happens to use FP 000.
2.6 Anycast Addresses
Anycast usage scenarios listed in the draft are very router-centric, and I
doubt they have been used at all (or very little).
Therefore, based on IPv6 anycast analysis draft, I think this would be
perfect time to remove/edit(move to recommendation or something) this:
o An anycast address must not be assigned to an IPv6 host, that
is, it may be assigned to an IPv6 router only.
Lest we be burdened for this for another 3 years or whatever, if e.g. a
DNS server wants to use anycast addresses.
The source address restriction is trickier business...
2.7 Multicast Addresses
FF01:0:0:0:0:0:0:101 means all NTP servers on the same interface
(i.e., the same node) as the sender.
FF02:0:0:0:0:0:0:101 means all NTP servers on the same link as
the sender.
Perhaps it would make sense to point out the differences of:
- interface
- link
- node
wrt. scoping (ref to scoping draft?), especially on scenarios where
interface != link. (interface really only includes the loopback, I think).
I think "interface-local" is a bit of a misnomer, but I guess changing it
to "node-local" at this point may be too late.
This has been discussed IIRC before, but it may not exactly obvious if
some first-timer reads the draft.
2.8 A Node's Required Addresses
Editorial:
o It's required Link-Local Address for each interface.
Its.
Editorial:
[TRAN] reference should probably be updated to point at RFC2893, not
RFC1933.
There's some fluctuation in the casing of some special words, like 'the
unspecified address' and 'loopback address'. The preferred form should be
decided and the occurrences where they're being used,fixed.
--
Pekka Savola "Tell me of difficulties surmounted,
Netcore Oy not those you stumble over and fall"
Systems. Networks. Security. -- Robert Jordan: A Crown of Swords
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
