How to Configure Single-Area OSPF on Cisco (Step by Step)

Step 1 — Enable OSPF and advertise networks

Start the OSPF process and tell it which interfaces to run on. The process ID (1 below) is locally significant — it doesn't have to match between routers. Set a router-id explicitly so the router keeps a stable identity even if an interface goes down.

The network statement uses a wildcard mask (the inverse of the subnet mask) and an area. Every interface whose IP falls inside a network statement runs OSPF and is advertised into that area.

R1(config)# router ospf 1
R1(config-router)# router-id 1.1.1.1
R1(config-router)# network 10.0.12.0 0.0.0.3 area 0
R1(config-router)# network 192.168.1.0 0.0.0.255 area 0

Step 2 — Configure the neighbor the same way

Every router in the domain needs the same treatment. On the link between R1 and R2, both sides must be in the SAME area (area 0 here) with matching subnet masks, or the adjacency will never reach FULL.

R2(config)# router ospf 1
R2(config-router)# router-id 2.2.2.2
R2(config-router)# network 10.0.12.0 0.0.0.3 area 0
R2(config-router)# network 192.168.2.0 0.0.0.255 area 0

Step 3 — Verify the adjacency and routes

Confirm the neighbor reached FULL, then check that OSPF-learned routes (marked O) are in the table. show ip ospf interface brief confirms which interfaces are actually running OSPF.

R1# show ip ospf neighbor
R1# show ip route ospf
R1# show ip ospf interface brief

A complete two-router walk-through (with the output to expect)

Concrete addresses make the config click. Picture R1 and R2 joined by a /30 transit link and each fronting its own LAN. The transit is 10.0.12.0/30 on Gi0/0 (R1 = .1, R2 = .2); R1's LAN is 192.168.1.0/24 on Gi0/1 (.1), and R2's LAN is 192.168.2.0/24 on Gi0/1 (.1). Router-ids are 1.1.1.1 and 2.2.2.2. Both LAN and transit subnets go into area 0.

Notice the one nuance the earlier 'Common problems' note hints at: mark the LAN interface passive so OSPF advertises that subnet but never wastes hellos on a segment full of hosts, yet leave the transit interface active, because making the R1-R2 link passive would silently prevent the adjacency from ever forming. That single distinction — advertise the LAN, speak on the transit — is where a lot of first attempts quietly break.

After both routers are configured, verification tells the whole story. The neighbor reaches FULL, and because Gi0/0 is Ethernet (OSPF broadcast network type by default) you'll see a DR/BDR role next to it even though only two routers share the link — R2 wins DR here on the higher router-id. The route to 192.168.2.0/24 arrives with an administrative distance of 110 and a cost of 2 (cost 1 across the Gig transit plus cost 1 for R2's LAN interface), which is exactly what the [110/2] in the table means.

! ----- R1 -----
R1(config)# interface Gi0/0
R1(config-if)#  ip address 10.0.12.1 255.255.255.252
R1(config)# interface Gi0/1
R1(config-if)#  ip address 192.168.1.1 255.255.255.0
R1(config)# router ospf 1
R1(config-router)#  router-id 1.1.1.1
R1(config-router)#  network 10.0.12.0 0.0.0.3 area 0
R1(config-router)#  network 192.168.1.0 0.0.0.255 area 0
R1(config-router)#  passive-interface Gi0/1     ! advertise the LAN, no hellos to hosts

! ----- R2 (mirror image) -----
R2(config)# router ospf 1
R2(config-router)#  router-id 2.2.2.2
R2(config-router)#  network 10.0.12.0 0.0.0.3 area 0
R2(config-router)#  network 192.168.2.0 0.0.0.255 area 0
R2(config-router)#  passive-interface Gi0/1

! ----- Verify on R1 -----
R1# show ip ospf neighbor
Neighbor ID     Pri   State      Dead Time   Address      Interface
2.2.2.2           1   FULL/DR    00:00:35    10.0.12.2    GigabitEthernet0/0

R1# show ip route ospf
      192.168.2.0/24 [110/2] via 10.0.12.2, 00:03:11, GigabitEthernet0/0

How OSPF picks the path: cost, and the reference-bandwidth trap

OSPF has no notion of hop count. Its metric is cost, and the cost of a route is simply the sum of the outgoing interface costs along the path to the destination. Each interface's cost is calculated as a reference bandwidth divided by the interface's bandwidth, and the classic default reference bandwidth on IOS is 100 Mbps (10^8 bps). That formula is why the worked example above showed a cost of 2 — one for each Gigabit-and-above interface the route crossed.

Here's the trap that bites people in modern networks: with the default 100 Mbps reference, any interface at 100 Mbps or faster floors out at cost 1. A GigabitEthernet link, a 10-Gig link, and a 40-Gig link all look identical to OSPF, so it can't prefer the faster path and may load-share across links you'd never want treated as equal. The fix is 'auto-cost reference-bandwidth' set to (at least) your fastest link speed in Mbps — and it must be set to the same value on every router in the domain, or the routers will compute inconsistent costs and choose different paths.

When you need surgical control rather than a domain-wide policy, override a single link with 'ip ospf cost' directly on the interface; a lower number is more preferred. Confirm whatever you set with 'show ip ospf interface', which prints the network type and the cost OSPF is actually using. One thing tuning cost never changes is OSPF's administrative distance — that stays 110, so cost decides between competing OSPF paths, while AD decides whether OSPF loses out to a lower-AD source like a static route or EIGRP.

R1(config)# router ospf 1
R1(config-router)# auto-cost reference-bandwidth 10000
% OSPF: Reference bandwidth is changed.
        Please ensure reference bandwidth is consistent across all routers.
R1(config-router)# exit
R1(config)# interface Gi0/0
R1(config-if)#  ip ospf cost 5           ! pin this link's cost by hand

R1# show ip ospf interface Gi0/0 | include Cost
  Process ID 1, Router ID 1.1.1.1, Network Type BROADCAST, Cost: 5

Common problems (and the fix)

If the adjacency won't form, check that both interfaces are in the same area, have matching subnet masks, and matching hello/dead timers; that neither interface is passive (passive-interface stops hellos); and that MTU matches on both sides. If routes are missing, make sure the network statement's wildcard mask actually covers the interface's IP.