Re: [Autoconf] RFC 5889 (Was: Call for comments to a new AUTOCONF charter proposal)

[Teco Boot <[e-mail]>]

Op 9 jul 2010, om 16:15 heeft Henning Rogge het volgende geschreven:
> 
>> And in Section 6.1:
>> 
>> OLD:
>>  o  There is no mechanism to ensure that IPv6 link-local addresses are
>>     unique across multiple links, hence they cannot be used to
>>     reliably identify routers (it is often desirable to identify a
>>     router with an IP address).
>> NEW:
>>  o  There is no mechanism to ensure that IPv6 link-local addresses are
>>     unique across multiple links, hence they cannot be used to
>>     reliably identify routers, should this be necessary by the routing
>>     protocol for which IP address autoconfiguration is being provided.
> 
> I think this is still too strong.
> "There is no mechanism to ensure..." does sound as there never will be one. If 
> we cannot find a mechanism/protocol to do this, we cannot find one that 
> creates routing unique adresses (because the later would do the former too).
> 
> I think it should start with something like this:
> "The default mechanisms of IPv6 cannot ensure that..."
> (instead of "There is no mechanism to ensure that...")

Although I think the document should not kept waiting, I don't agree on this.
There is no mechanism to detect duplicate globals on different links.
The ND DAD mechanism does not work in a MANET for LL, and _also_ not for globals !!
A mechanism that ensures unique Interface-IDs ensures unique LLs.

We all agreed on that we shall not work on a new protocol for autoconfiguring LLs.
It is globals and MANET-locals (ULAs) that we are looking for.
That is in the document, in old and new text.

I would use unique LLs as router-IDs. I don't get what is wrong with this, as long as LLs are unique.
But LLs MUST NOT be used for multi-hop forwarded traffic.

Teco.