bgp behavior with exchange prefix in ebgp

Hi,

On Thu, Oct 02, 2003 at 10:10:51AM +0200, Havard Eidnes wrote:
> >  - filter the exchange prefix(es) for all IXes that you participate *and 
> >    all more specifics* on *all* eBGP sessions.  This is strongly recommended
> >    even when already using next-hop-self (more specifics can break eBGP
> >    sessions at the IXP peering routers)
> 
> A third and additional defensive measure could be to adjust the
> preference of your IGP to a value below that for eBGP, and do the
> moral equivalent of "redistribute connected" on the exchange
> point router into your IGP.

While I agree that it's useful, it still needs measure 2 - the problem
being more-specific annonucements of the IXP prefixes (which can't be
caught by the IGP<->BGP preference).

Gert Doering
        -- NetMaster
-- 
Total number of prefixes smaller than registry allocations:  56883  (56833)

SpaceNet AG                 Mail: netmaster at Space.Net
Joseph-Dollinger-Bogen 14   Tel : +49-89-32356-0
80807 Muenchen              Fax : +49-89-32356-299