We have a situation. The BGP routing table size is too large in the Router pair DR1/2, that has only one (redundant) path(s) to internet router iR1/2 pair.
On DR1 and 2, I'm planning to discard BGP route updates from peers iR1/2, and instead redistribute default route (pointing to iR1/2) on the DR1/2 into BGP.
Is this good approach?
Also how will iR1/2, without any change, handle the private IP subnets, as the DR1/2 might send private traffic, because of default route?
(Please see attached .doc)
Thanks for checking this
if DR1/2 cannot support a full table you need to filter and you should do it directly on IR1/2
ip prefix-list only-default seq 5 permit 0.0.0.0/0
route-map toDR permit 10
match ip address prefix only-default
neigh dr1 route-map toDR out
neigh dr2 route-map toDR out
note: filtering on DR1/2 inbound is inefficient because IR1/2 will try to advertise all prefixes just to have them dropped on DR1/2
as a security measure on DR1/2 you can implement the same filter inbound
on DR1/2 you keep to advertise the private ip subnets in order to ensure correct return traffic
this helps in having a dynamic routing between the 4 routers
I would not use static routes otherwise you will need to use IP SLA to control if next-hops are alive in a LAN environment (with switches in the middle)
Hope to help