Re: [Roll] security for multi-link subnets

[Don Sturek <[e-mail]>]

Hi Michael (and Ulrich),

The area that I think needs more work around the multi-link subnets topic
is the scope for site-local addresses.

Since link locals are not that useful in multi-link subnets (routing wise)
and not all devices in a 6LoWPAN/ROLL environment may want global
addresses even if such a prefix were available, clear scoping rules (eg,
routing, multicast propagation) on site local addresses is a major topic.
If someone knows of an RFC (or even a draft) that starts to address the
issue of identifying site local versus non-site prefixes/interfaces, that
would be interesting.  This topic is especially interesting in deployments
with multiple prefixes (like that now being discussed in Homenet) as well
as campus type environments.

I also bring this up (on purpose) under the topic of "security for
multi-link subnets" since turning on link security does not help securing
these networks.......

Don



On 3/12/13 12:46 PM, "Michael Richardson" <[e-mail]> wrote:

>
>>>>>> "Ulrich" == Ulrich Herberg <[e-mail]> writes:
>    Ulrich> I think it is also worth mentioning RFC4903, in particular:
>
>    Ulrich> "A multi-link subnet model should be avoided.  IETF working
>groups
>    Ulrich> using, or considering using, multi-link subnets today should
>    Ulrich> investigate moving to one of the other models."
>
>    Ulrich> Have the issues mentioned in RFC4903 been sufficiently
>addressed?
>
>I think that if we were going supposed to avoid a multi-link subnet,
>that would have been objected to already.
>I think that 4903 concerns applied to all of 6lowpan and ROLL work, and
>I think that actually we did deal with all of these.
>
>-- 
>Michael Richardson
>-on the road-
>
>
>_______________________________________________
>Roll mailing list
>[e-mail]
>https://www.ietf.org/mailman/listinfo/roll