Re: [DNA-BOF] Draft BoF agenda.

[Greg Daley <greg.daley@eng.monash.edu.au>]

Hi Erik,

Erik Nordmark wrote:
[cut message]

>>
>>Then what is the Router1's on-link prefixes on the wireless link? 
>>In this case, is the complete set of on-link prefixes  {Prefix A::} 
>>or {Prefix A:: , Prefix B::}
> 
> 
> I think the key question is what prefixes are assigned to the link
> and in this case both A and B apply. The same would be the case of
> both routers advertised both A and B, or if there was a single router
> advertising A and B.
> 
> I think the confusion is caused by the attempt to overload on-link prefixes
> to mean something different than intended. THeir meaning is: you can and
> should perform address resolution for an address that falls in the on-link
> prefix and send packets addressed to other routers to one of the default
> routers.
> RFC 2461 allows links that have no on-link prefixes i.e. where all packets
> would be sent to a default router (with the ability to redirect to on-link
> destinations).
> 
> Thus as RFC 2461 is designed it is problematic to overload the on-link
> prefixes as some form of "link identifier".
> 
> I see two choices on how to proceed:
>  - revisit RFC 2461 and make all routers on the link advertise the same set
>    of on-link prefixes and require that at least one on-link prefix
>    be advertised on every link.
>  - introduce an explicit "link identifier" option in Neighbor Advertisements.

Brett, JinHyeock and I have elaborated on a mechanism
for a 'Link identifier' modified from your description in
the mipv6 hindsight draft.

At this stage, we kept the identifier to RAs, although it
certainly could also be present in NAs from Routers.

This document is part of the reading list for the BoF.
(draft-pentland-mobileip-linkid-00.txt)

For obvious reasons I think that updates to Router
Discovery would have to be backward compatible with RFC-2461.
Maybe the common on-link prefix is equivalent to a
Link Identifier option (since it cannot be universal).

Greg