Re: [Autoconf] Last Call: draft-ietf-autoconf-adhoc-addr-model (IP Addressing Model in Ad Hoc Networks) to Informational RFC

[Erik Nordmark <[e-mail]>]

On 03/24/10 03:37 PM, Alexandru Petrescu wrote:

> I read Erik's email as suggesting a decouple between the use of the
> link-local address and of a Router ID. (his "deflate" to mean contrary
> to "inflate"). It may be that the following paragraph in the draft led
> to think so:
>
> draft-ietf-autoconf-adhoc-addr-model:
>> 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).
>
> ... because whereas it _is_ indeed desirable to identify a router with
> an IP address, it is not necessarily the case that that IP address is
> the address of an interface. E.g. OSPF does so with a Router ID being
> an address (an IPv4 address), and different than any IPv6 addresses it
> may have on an interface (be them link-local addresses)
 >
> Why does the above paragraph suppose that the Router ID where the
> link-local address?

There is other text that conflate them such as this text in section 5
    o  an IP address assigned to an interface that connects to a link
       with undetermined connectivity properties should be unique, at
       least within the routing domain.
Similar text specific to IPv6 is in a later section.

The motivation for this seems to be to get a unique router ID across the 
domain, which I hopefully succeeded explaining is a separate issue from 
the IP addressed assigned to the router's interface(s).

>> BUT, the document 'suggest' not to use the link-local address for
>> routing protocols
>
> That suggestion is a little daring because it goes against deployed
> OSPF, RIP, RFC4191, ICMPv6, DHCPv6, and more.

Daring and inconsistent with existing practice.

>> and data packet forwarding.
>
> Forwarding an ll-addressed packet across _known_ links should be indeed
> forbidden. Draft says so.
 >
> Minor aspect: it doesn't say what's a link other than having
> undetermined connectivity properties, i.e. we don't know what is a link,
> hence get rid of Link-local addresses.
>
> This reason of undetermined link properties is a little difficult to
> grasp for me.

I thought the WG had discussed this for a few years. I'm sorry if it is 
still not clear.

    Erik