OSPF

While RIP is easy enough to implement and use in small networks, it becomes a problem if the network becomes larger. The solution is the use of another protocol: OSPF

OSPF uses a different protocol than RIP, and to understand OSPF, a set of concepts must be clear.

For routing, there are two types of protocols:

• Distance Vector
• Link State

Distance vector protocols, like RIP, blurt out all the information that they know, and listen to any updates that others send. The information is processed following the Bellman-Ford agorithm.

1. Increment the merics of the incoming routes.
2. Compare the incoming with the existing routing table
3. if the neighbour's information is better: replace the old entry in the routing table
4. if the information is worse, discard it
5. if it is the same, just reset some timers
6. If it has the same costs, but a different path, add it to the routing table anyway

Or,

In case of RIP, it is easy to understand step 1: the neighbour advertises its own hop-count and to the neighbour, it is one hop extra.

Link State protocols, like OSPF, generally only generate messages when a topology changes. They use multicasts in stead of broadcasts and they send their updates reliably to other routers. The other routers will reply with an acknowledgement.

Link State routers will build a topology of the network, called the topology table. To calculate the best route, Dijkstra's SPF algorithm is used. Dijkstra's algorithm is a subject by itsself; I will content myself with a reference to Packet Pushers.

All routers in an OSPF network must have a unique ID (that is: unique in the OSPF network. The router ID is chosen according to one of the following criteria:

• Manually defined
• The highest IP address on the router’s active loopback interfaces
• If no loopback interface exists with an IP address, the highest IP address on its active interfaces

Routers on the borders of areas and Autonomous systems have special names:

Routers maintain contact with adjacent routers by sendin HELLO messages.

OSPF uses Link Stata advertisements to communicate the link state between routers. The most common LSA types are:

An Autonomous system is a part of the network that is under control of a single entity. IANA keeps a list of Autonomous systems that are publicly accessible, but as long as you don't connect to the Internet, you can use private numbers. IANA has reserved, for Private Use, a contiguous block of 1023 Autonomous System numbers from the “16-bit Autonomous System Numbers” registry, namely 64512 – 65534 inclusive. IANA has also reserved, for Private Use, a contiguous block of 94,967,295 Autonomous System numbers from the “32-bit Autonomous System Numbers” registry, namely 4200000000 – 4294967294 inclusive.

OSPF implements a two-level hierachy for areas:

• Area 0, which is the backbone
• All other areas

In general, all other areas are connected to the backbone.

We have now seen two types of area:

• backbone (area 0)
• standard area

But OSPF is a bit more complicated than that. There are also

• stub areas
• totally stubby areas
• not so stubby areas

and fairly stubby areas, extremely stubby areas, just a bit stubby areas, and ... No, I am exagerating.

A stub area is an area through which or into which AS external advertisements are not flooded. This means that you would only be able to access devices within the AS from this type of area. This is done to reduce the size of the topology database.

Totaly stubby areas takes this one step further by not even allowing LSA type 3; they are replaced by a default route.

Not So Stubby Areas allow external routes to be flooded within the area. These routes are leaked into other areas. However, external routes from other areas within the AS are not flooded into these areas.

You can view the area-types as whether they accept certain types of LSA: