Re: [DNA] Proposed resolution to issue 9 fordraft-pentland-dna-protocol-01.txt

[Syam Madanapalli <smadanapalli@gmail.com>]

I was thinking that the completeRA may not be complete all the time.
If the 'complete RA' is assumed then it is ok to check the prefixes in
the RA received.

-Syam

On 8/3/05, Syam Madanapalli <smadanapalli@gmail.com> wrote:
> Hi Brett,
> 
> I think we may need to check for the prefixes that the other routers
> on the link advertising. That is we need to build CPL on the new
> link.
> 
> -Syam
> 
> 
> On 8/3/05, Brett Pentland <brett.pentland@eng.monash.edu.au> wrote:
> > Hi DNA WG,
> >
> > All of the recent discussion on the issue of RAs that include
> > a "no" answer in a Landmark Option but also include prefixes
> > that have been seen on the link, has been in favour of
> > treating this as an indication that link change has not
> > taken place.  Here is some proposed text to reflect this:
> >
> > Old text:
> >
> >       Instead if the 'N' bit is set in the received Landmark Option, a
> >       change of link is indicated and the host SHOULD initiate
> >       reconfiguration using the information in the Router Advertisement.
> >
> > New text:
> >
> >       Instead if the 'N' bit is set in the received Landmark Option, AND
> >       there is no overlap between the list of prefixes in the Router
> >       Advertisement and the list of prefixes that the host has stored
> >       for CPL [14], a change of link is indicated and the host SHOULD
> >       initiate reconfiguration using the information in the Router
> >       Advertisement.
> >
> >       If the 'N' bit is set in the received Landmark Option, but there
> >       is one or more prefixes in the Router Advertisement that are also
> >       in the list of prefixes the host has stored for CPL, this is an
> >       indication that no link change has occured but there is prefix
> >       change underway on the link.  The host MUST NOT invalidate its
> >       current configuration.
> >
> > Any comments?
> > Brett.
> >
>