CCIE - Internetwork Expert's Online Community

• Configure the network in such a way that these routers still have reachability to all BGP learned prefixes, but do not need to carry a full view of the Internet  routing table.

There are a lot of drawbacks associated with SG's solution. I decided to solve this task using redistribution.

Rack1R6(config-router)#do sh run | s eigrp 10
 router eigrp 10
 redistribute bgp 100 metric 1 1 1 1 1 route-map BGP->EIGRP

Rack1R6(config-router)#do sh run | b pref
ip prefix-list BGP->EIGRP seq 5 permit 28.119.16.0/22
ip prefix-list BGP->EIGRP seq 10 permit 112.0.0.0/5
!
!
!
route-map BGP->EIGRP permit 10
 match ip address prefix-list BGP->EIGRP

Rack1R5(config-router)#do sh run | b pref
ip prefix-list BGP->EIGRP seq 5 permit 28.119.16.0/22
ip prefix-list BGP->EIGRP seq 10 permit 112.0.0.0/5
ip prefix-list BGP->EIGRP seq 15 permit 205.90.31.0/24
ip prefix-list BGP->EIGRP seq 20 permit 220.20.3.0/24
ip prefix-list BGP->EIGRP seq 25 permit 222.22.2.0/24
!

route-map BGP->EIGRP permit 10
 match ip address prefix-list BGP->EIGRP

router eigrp 10
 redistribute bgp 100 metric 1 1 1 1 1 route-map BGP->EIGRP
 

This way we don't violate task req and in the same time provide optimal and working solution.

My issue with the SG here is that it break section 4.3 as well.  The leak-map needs to include the 200.0.0.0/22 and 54.1.1.0/24 network that are learned from R6 / BB1.  I just added another sequence to leak-map on R5 to address this.