If you are just looking into routing protocols for use in your network, we talked about EIGRP in a previous article. However, don’t just go implement it without a little comparison to other options. If you research only one other routing protocol, I would highly recommend OSPF. OSPF is standards based, which is one of the attractions of this protocol. This means different vendors network routing equipment can share routing tables by using OSPF. Unlike EIGRP, you don’t have to have all Cisco equipment in your network. Other network components (Cisco or not) can also benefit from using OSPF, such as Firewalls and VPN concentrators. This prevents having to use static routing to reach hosts on the other side of these types of components.

OSPF can be as simple or as complicated as one wants to make it. In this example, we will enable OSPF in a simple network that one might find in small office environments.

First let’s start with a basic network containing three routers. The Lan and Wan Routers are in the same ‘Main’ office with the BranchRouter located at a secondary location. Even though this network could be routed easily with static routes, we want to go ahead and start with a routing protocol to be prepared for future expansions..

hostname LanRouter
!
interface FastEthernet0/0
description Infrastructure Subnet
ip address 10.0.1.1 255.255.255.0
!
interface FastEthernet1/0
description Employee Subnet
ip address 10.0.2.1 255.255.255.0

hostname WanRouter
!
interface FastEthernet0/0
description Infrastructure Subnet
ip address 10.0.1.2 255.255.255.0
!
interface Serial1/0
description Point to Point link to Branch Office
ip address 192.168.1.1 255.255.255.252

hostname BranchRouter
!
interface FastEthernet0/0
description Branch Employee Subnet
ip address 10.1.2.1 255.255.255.0
!
interface Serial1/0
description Point to Point link to Main Office
ip address 192.168.1.2 255.255.255.252

If we want to route between these two offices, here is all the routing configuration that is needed:

LanRouter(Config)#router ospf 1
LanRouter(Config)#network 10.0.1.0 0.0.0.255 area 0
LanRouter(Config)#network 10.0.2.0 0.0.0.255 area 1

WanRouter(Config)#router ospf 1
WanRouter(Config)#network 10.0.1.0 0.0.0.255 area 0
WanRouter(Config)#network 192.168.1.0 0.0.0.3 area 2

BranchRouter(Config)#router ospf 1
BranchRouter(Config)#network 10.1.2.0 0.0.0.255 area 2
BranchRouter(Config)#network 192.168.1.0 0.0.0.3 area 2

The big difference between the OSPF configuration shown above and the EIGRP configuration in the other article is the ‘area’ keyword. OSPF uses areas to segment the routing protocol throughout your network. If a change in the network occurs, all areas do not recalculate automatically. It is the responsibility of certain OSPF routers in the network to determine whether of not some areas need any type of notification. This behavior is meant to improve the efficiency of the routing protocol and improve performance of things such as link failover times.

All OSPF networks are required to have an area 0. In the example above; area 0 is placed in the ‘core’ of network, which is the Main Office, specifically the 10.0.1.0 subnet. This network is the shared subnet between the Lan and Wan routers, which is where area 0 should be placed in this example. Area 1 is configured for the local subnet(s) used at the Main Office for PCs. Area 2 is configured for the Branch Router network, which includes the serial subnet used between the WanRouter and BranchRouter.

Any router that has an interface in Area 0 and at least one other area is called an Area Border Router (ABR) in OSPF. In this example, the LanRouter and WanRouter are both ABRs. These two routers have the responsibility mentioned earlier of determining whether routers in other areas (outside of area 0) need notification to recalculate their routing table.

With all of this said, a network this small can have all of the router interfaces in area 0. This will work just fine, the downfall is as the network grows - every router will be doing re-calculations anytime a network change occurs. There are much larger networks that exist in production environments today using nothing but Area 0, but it is very inefficient and does not provide for optimal failover response times, among other things.