EIGRP and OSPF PDF Print E-mail

EIGRP or Enhanced Interior Gateway Routing Protocol is a hybrid routing protocol. It is called a hybrid because it has the characteristics of a distance-vector and link-state protocols. EIGRP was developed for use in bigger networks, its maximum hop count is 255 hops. Other characteristics of EIGRP that makes it more powerful are, support for IP, IPX and Apple talk, efficient discovery of neighbors, support for VLSM AND CIDR and it is also considered classless.

Having support for multiple protocols make EIGRP very useful, but it also makes it more complex. PDM’s are protocol dependent modules, each module keeps a separate table for each specific protocol.

Hello messages are very important for EIGRP routers, they use Hello messages to start a neighbor relationship with routers, and also use them to maintain this relationship. Hello messages are not the only conditions for a neighbor relationship to begin. First the autonomous system numbers must match, this means that EIGRP routers that belong to different AS numbers don’t automatically share routing information. You would you think that this would be a bad thing, but in larger networks this reduces the amount of routing information propagated throughout the network. Lastly a network must have identical metrics or K values. When these three conditions are met a neighbor relationship can begin.

Now you might as what is a AS number? Well an AS number is a number used to identify which collection of routers share information AS numbers can be used to break a complex topology into a rather simple collection of routers.

So how do these routers share information and all these messages that we are talking about? Reliable Transport Protocol or RTP is used for the communication between EIGRP routers. A router keeps track of its neighbors by sending out multicast, it then keeps track of all neighbors that respond, for each neighbor that does not respond it sends out a unicast, after 16 non-replies the router declares the neighbor dead.

So now we know how the routers discover each other, but how do they store this information? Each router has a neighbor table, a topology table and a route table. These tables store the information for each router. The Neighbor table keeps the information about adjacent routers. The topology tableuses a Diffusing Update Algorithm or DUAL to choose which route is best.

One of the best things about EIGRP is that it uses four different factors to determine the best route for a packet. It uses Bandwidth, Delay, Load and Reliability to determine the best route. This helps speed up convergence in a topology. I have said that EIGRP is good for bigger networks, because of speed, communication, reliability, and its versatility in regards to other protocols. But how does a network administrator have enough IP addresses for this large network we have been talking about? Well EIGRP uses something called VLSM, which stands for Variable Length Subnet Masks. This allows for the conservation of IP addresses by using subnet masks. OSPF What do you do if not all the routers on your network are Cisco? Well then you use a protocol called OSPF, or Open Shortest Path First. OSPF is a translation service between routing protocols such as IGRP. A administrator would use OSPF on larger networks, but would use RIP routing on smaller networks. OSPF has many characteristics, but one of its best is that it allows a network to be created with a hierarchical design. This means a network can be broken down from one larger internetwork to many smaller internetworks. Why you might ask would I want to make it more complicated for myself? Well it decreases routing overhead which speeds up convergence. Also if there is a problem the network instability is confined to one area. OSPF uses Areas to group each internetwork. How does OSPF compute routes you ask? OSPF uses an algorithm. The Dijkstra algorithm works like this, First, a shortest path tree is created, this then inputs the shortest paths into a routing table. This is called SPF or Shortest Path First.

*** Put in OSPF Hierarchical design figure

Several other features of OSPF are that it has  unlimited hop count, scalablity, and support for VLSM/CIDR.

When there are changes in the network, OSPF uses a multicast, to find out about the changes. With all these features OSPF is much more complex than its understudies RIPv2 and RIPv1. Each RIP protocol is much easier to setup, but convergence is much faster with OSPF. And being a network administrator your company will appreciate this. One drawback to OSPF is that it only supports IP routing.

Now that we have covered these two protocols we need to cover some terminology that you will come across.

First we will cover OSPF, but remember, these terms will be used in other areas besides OSPF.

Neighbors: Two routers connected by an interface on a common network. Hello Protocol: Hello packets and LSA’s or Link State Advertisements build and maintain a topology database for each router.

Link: A router or a network interface are considered a link.

Neighbor Database: When the hello packets are sent, the results are stored in the neighbor database.

Topology Database: When Link State Advertisement Packets are received they are put into the topology database where the Dijkstra algorithm computes the shortest path. To further this definition LSA’s are packets that contain information about the routers.

DR & BDR- Designated router, the primary, and Backup designated router, the secondary router. The Designated router broadcast all routing information on the network. The backup is on hot standby in case the DR goes down. The DR is chosen because it has a higher priority.

An Area is a grouping of routers and networks, designated under a specific number.

A Broadcast is a single cloned packet sent to every router on the network.