OSPF supports a number of methods to filter routes but it is more restrictive compared to distance vector routing protocols like RIP or EIGRP.
As a link-state routing protocol OSPF uses LSAs to build its LSDB (Link State Database). Routers will run the SPF algorithm to find the shortest path to each destination, the topology in the LSDB has to be the same on all routers or SPF will fail.
However OSPF routers only know what the topology looks like within the area. They don’t know what the topology looks like for other areas. For inter-area routes OSPF only knows the prefix and the ABR (Area Border Router) to reach it.
You could say that OSPF acts like a distance vector routing protocol for inter-area routes. It only knows the metric (distance) and the ABR to get there (vector).
Unlike RIP or EIGRP, OSPF doesn’t advertise routes but LSAs so if we want to filter something we’ll have to filter the advertisement of LSAs.
Since the LSDB within the area has to be the same we can’t filter LSAs within the area, we can however filter routes from entering the routing table. Filtering LSAs between areas on an ABR or ASBR is no problem.
Here’s the topology I will use:
Nothing fancy, we have three routers running OSPF in the same area. R1 has a loopback interface that is advertised in OSPF, we’ll see if we can filter this network.
Here’s the OSPF configuration:
R1#show running-config | section ospf router ospf 1 network 126.96.36.199 0.0.0.255 area 0 network 192.168.12.0 0.0.0.255 area 0
R2#show running-config | section ospf router ospf 1 network 192.168.12.0 0.0.0.255 area 0 network 192.168.23.0 0.0.0.255 area 0
R3#show running-config | section ospf router ospf 1 network 192.168.23.0 0.0.0.255 area 0
Let’s verify if R2 and R3 have learned 188.8.131.52 /32:
R2#show ip route ospf 184.108.40.206/32 is subnetted, 1 subnets O 220.127.116.11 [110/2] via 192.168.12.1, 00:00:27, FastEthernet0/0
R3#show ip route ospf 18.104.22.168/32 is subnetted, 1 subnets O 22.214.171.124 [110/3] via 192.168.23.2, 00:00:28, FastEthernet0/0 O 192.168.12.0/24 [110/2] via 192.168.23.2, 00:00:28, FastEthernet0/0
Let’s see if we can get rid of this network on R3:
R3(config)#router ospf 1 R3(config-router)#distribute-list ? <1-199> IP access list number <1300-2699> IP expanded access list number WORD Access-list name gateway Filtering incoming updates based on gateway prefix Filter prefixes in routing updates route-map Filter prefixes based on the route-map
We can use a distribute-list for this, to keep it simple I’ll combine it with an access-list;
R3(config-router)#distribute-list R1_L0 in
When we want to remove something from the routing table we have to apply it inbound. The outbound distribute-list is used for LSA type 5 filtering.
Let’s create that access-list:
R3(config)#ip access-list standard R1_L0 R3(config-std-nacl)#deny host 126.96.36.199 R3(config-std-nacl)#permit any
It will now be gone from the routing table:
R3#show ip route 188.8.131.52 % Network not in table
As you can see it’s gone…it’s still in the LSDB though:
R3#show ip ospf database router 192.168.12.1 OSPF Router with ID (192.168.23.3) (Process ID 1) Router Link States (Area 0) LS age: 664 Options: (No TOS-capability, DC) LS Type: Router Links Link State ID: 192.168.12.1 Advertising Router: 192.168.12.1 LS Seq Number: 80000003 Checksum: 0xF14F Length: 48 Number of Links: 2 Link connected to: a Stub Network (Link ID) Network/subnet number: 184.108.40.206 (Link Data) Network Mask: 255.255.255.255 Number of MTID metrics: 0 TOS 0 Metrics: 1 Link connected to: a Transit Network (Link ID) Designated Router address: 192.168.12.2 (Link Data) Router Interface address: 192.168.12.1 Number of MTID metrics: 0 TOS 0 Metrics: 1
You have to be very careful if you use this command. If you are not careful you can end up in a scenario where you blackhole some traffic. For example, let’s see what happens when I filter this network on R2 instead of R3.