[nsp] Methods for Non-BGP multihoming

[Ray Davis]

I guess the author wasn't really sure either:

    "Generally, a prefix can should belong to only one AS."
			 ^^^^^^^^^^
:)

The statement in the RFC assumes that an IP address exists in only one
place.  This isn't always true.

Probably the statement should be removed from the RFC since it doesn't
really say anything firm, gives no good reason and says it can and may
be otherwise anyway.

Cheers,
Ray

> Actually, RFC 1930 has a more direct reference regarding inconsistent AS
>
> http://asg.web.cmu.edu/rfc/rfc1930.html#sec-7
>
>
> > -----Original Message-----
> > From: Ryan O'Connell [mailto:ryan@complicity.co.uk]
> > Sent: Friday, August 02, 2002 7:35 AM
> > To: Daniel Golding; cisco-nsp@puck.nether.net; joe.provo@rcn.com
> > Subject: Re: [nsp] Methods for Non-BGP multihoming
> > 
> > 
> > On Wed, 31 Jul 2002 14:05:49 -0400 Daniel Golding 
> > <dgolding@sockeye.com> wrote:
> > > RFC 1771 does indeed specify that inconsistant origin AS is an error
> > > state.
> > 
> > I just reread the RFC and I can't see that anywhere.
> > 
> > In fact, origin AS information is lost when AS_SEQUENCEs are 
> > converted to
> > AS_SETs from what I can tell so it would be difficult to 
> > enforce this anyway.
>
> > 
> _______________________________________________
> cisco-nsp mailing list  real_name)s@puck.nether.net
> http://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/