OSPF Forwarding Address

Hey All - I've got a scenario presented by Cisco Live CCIE R&S that I can't figure out how they can actually make this work. The link is here:

https://learningnetwork.cisco.com/docs/DOC-24192 (check the slide at 2:38)

R1,R2, R3 are on a shared Eth segment, all interfaces have an ospf cost of 10. R2 - R4, cost 10, R3 - R5, cost 10. All OSPF area 0. Then R4,R5, and R6 are on another shared Eth Segment, not running OSFP. R4 and R5 are redistributing static routes into OSPF with a default metric.

Then R5 adds the interface into OSPF, creating an LSA with the forwarding address bit set.

Now I completely understand why R1 will choose the LSA from R4 (no f.w. address) over R5's.

HOWEVER: the "current situation" states the following:

- R1 installs only one OSPF external router for the destination 200.200.200.0/24 (from R6) while it learns two OSPF LSA type 5 for which the metric to the ASBR is equal
(I get this)
- R2 installs two OSPF external routes for the same destination while it's metrics to both ASBR's are not equal!
(I can not figure out how to do this)
- R3 installs only on OSPF external route for the same destination while it learns two OSPF LSA type 5
- Both R4 and R5 have a static route for the same destination
- R6 installs two equal cost default static routes towards the OSPF domain, via R4 and R5.

So on bullet #2 - how can R2 have 2 equal cost routes? It is comparing the SPT cost to reach the ASBR with no f.w. address, to the metric of the route of the f.w. address. If all the costs are 10, I don't see how these two comparative costs can be equal?

Blogs and organic groups at http://www.ccie.net
Received on Sun Sep 07 2014 - 19:10:37 ART