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

[Alexandru Petrescu <[e-mail]>]

Le 21/07/2010 11:58, Emmanuel Baccelli a écrit :
> Hi Erik,
>
>
> On Fri, Jul 9, 2010 at 10:52 PM, Erik Nordmark <[e-mail]
> <mailto:[e-mail]>> wrote:
>
>     ...
>
>     The result is that the document can easily be construed as
>     discouraging approaches that make a lot of sense, which seems
>     counterproductive.
>     An example of such an approach is a routing protocol which uses IEEE
>     MAC addresses as router ids, assigns IPv6 link-local addresses to
>     the router's interfaces and uses that in the routing protocol
>     exchanges, and configures global addresses for use by applications.
>
>
> Using MAC addresses for unique identifier is considered dangerous by
> many people, as MAC addresses are not always unique. There were several
> cases where an order of several thousands of devices where shipped with
> the exact same MAC address, by mistake. Moreover, MAC addresses can be
> altered almost at will. This is why relying on MAC addresses for unique
> identifiers is not sufficient in my mind. It's merely another way to
> say: "let other people worry about our problems". This rarely works
> reliably.
>
> As to using link local addresses, there are many cases where it is not
> appropriate,

Sure, and there are many cases where they are appropriate.  Why 
mentioning only the negative part?

> as discussed over and over in this working group and
> elsewhere. If you have any comments on
> http://ietfreport.isoc.org/all-ids/draft-baccelli-multi-hop-wireless-communication-04.txt
> please let us know. The fundamental properties of multi-hop wireless
> communications described in that document make it difficult to use
> link-local addresses in many cases.

Hmmm... sure, in that particular draft deployment.

In the multi-hop wireless communications system I build the link-local 
addresses are very useful and actually can't work without them.

Alex

>
> And again, the goal of RFC5889 is not to describe all possible
> addressing models for each specific scenario, but rather to describe one
> specific model with which we have experience in multi-hop wireless
> contexts, and which has proven to work OK. In that respect, I
> think RFC5889 does the job.
>
> Regards,
> Emmanuel
>
>
>
>
> _______________________________________________
> Autoconf mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/autoconf