Hi, We use route servers to mesh all the PEs. On the route server I can see that all the prefixes (even without the RT, just with RD) are learnt (or rather received), that causes unnecessary traffic amongst the route servers. The direct peering between the PEs is done within the vrfs - ie each PE has a loopback in a vrf Internet and the ibgp sessions directly on those loopbacks. Those sessions only exchange the 'internet' prefixes and export default into other vrf, so basically it works like if they were completly independed routers.
kind regards Pshem _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
