>>>>> On Tue, 12 Nov 2002 13:53:00 +0100,
>>>>> Brian E Carpenter <[EMAIL PROTECTED]> said:
> Unfortunately it's too late to catch the addressing architecture
> document unless we recall it from the RFC Editor and cycle it
> through the IESG again. But I propose that we do exactly that,
> in order to change the following paragraph in section 2.5.6:
> Current text:
>> Site-local addresses are designed to be used for addressing inside of
>> a site without the need for a global prefix. Although a subnet ID
>> may be up to 54-bits long, it is expected that globally-connected
>> sites will use the same subnet IDs for site-local and global
>> prefixes.
> Proposed new text:
> Site-local addresses are designed to be used for addressing inside of
> a site which is not connected to the Internet and therefore does not
> need a global prefix. They must not be used for a site that is connected
> to the Internet. Using site-local addresses, a subnet ID may be up to
> 54-bits long, but it is recommended to use at most 16-bit subnet IDs,
> for convenience if the site is later connected to the Internet using a
> global prefix.
I'm personally okay to restrict the use of site-local addresses
(though the level of the restriction may need to be discussed).
However, the proposed text seems to introduce a big impact on the
current usage of SLs in spite of the small change on wording; it
restricts the use to non-connected networks. This type of proposal
has been raised in this thread several times, and we've seen doubts
and objections to the proposal every time (of course, as well as
agreement).
So, I'm a bit surprised that a majority of this group is going to
support the proposed text, even by those who had a doubt such as "I
personally would like to deprecate site-local addresses, but I'm not
sure if it is effective to impose such a big change on the current
usage."
I admit I've not read all the messages in this huge thread, but please
let me ask: has there been a consensus in the thread to limit the use
of site-locals to not-globally-connected networks? If so, I'll follow
the consensus, and will just be happy. Otherwise, I'd say we have to
be careful because we'll easily go into the same loop of discussion.
I'd also like to point out a few things (some of them have been
covered by the discussion so far):
- the case of multi-sited nodes, which is a major sources of
headaches, is not eliminated even if we can restrict the usage to
non-globally-connected networks. What should we do on this topic?
- though I'm okay to restrict the use of SLs (as I said above), I'm
not sure if the proposed limitation is effective. What exactly does
the "must not" mean? Does it mean nodes in the global network must
drop packets contains SLs? How can we ensure the restriction? How
can the nodes detect if they are in the global network? What should
we do when our network using SLs connects to the global network?
Having considered the points above, my honest feeling is that the
proposal is an incomplete compromise. The impact is as big as a
stronger change of deprecating SLs *completely*, but there will still
remain difficult issues. If we have really reached a consensus on the
proposed text, I guess we can even agree on removing SLs completely,
with which I'll be much happier.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
