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 03/26/10 11:04 AM, Jari Arkko wrote:
> Erik,
>
> First of all, the document is *an* addressing model for ad hoc networks.
> It does not claim to be the only model. For instance, during working
> group discussions it also became apparent that link local addresses
> could also be employed, albeit -- as stated in the document -- the
> working group prefers other types of addresses.
>
> Note that this document talks about prefix and address configuration,
> not router IDs. You assume that we talk about router IDs because the
> document says that it would be beneficial to have the addresses be
> unique within a routing domain. But that's not the reason. The reason is
> that in an ad hoc network you may end up being a neighbor to any other
> device, and you want to avoid an address collision merely for the
> reasons of not getting a collision of two same addresses in the same link.

This paragraph talks about the requirements of the routing protocols, 
not about the need for routers to move around:
    Routing protocols running on a router may exhibit different
    requirements for uniqueness of interface addresses; some have no such
    requirements, others have requirements ranging from local uniqueness
    only, to uniqueness within, at least, the routing domain (as defined
    in [RFC1136]).

The only requirement I know if in the *routing protocols* is around 
router ID uniqueness.
If you disagree, then can you please explain why there is a need to 
refer to routing protocols in the above paragraph?

    Erik