Quick Guide to OSPF

► 1. Building Shortest Path Tree Topology

    2. Redistributing External Network Routes into OSPF Network

 

This time, we will discuss how routers discover IP network topologies and build shortest path trees by using OSPF protocols.

Certainly, routing protocols like OSPF are based on operating principles that are hard to understand if you are not a routing expert. Many people are quite good at routing configuration using Command Line Interface (CLI), but only a few fully understand the operating principles of the protocols. To help you, routing novices, understand OSPF better, below we have provided lots of easy-to-understand illustrations.  

 

 

OSPF network topology

 

 

In the figure, there are five OSPF routers (R1 ~ R5), along with their respective interface IP address, OSPF cost and loopback address (e.g. 10.1.0.1 for R1).

• Here OSPF cost means cost needed to deliver a packet to each link. From OSPF's perspective, a least cost path is the shortest path. In general, the cost of a link is determined based on its bandwidth. For example, the cost of 100Mbps link is greater than that of 1GE link, and the cost of 1GE of greater than that of 10GE (The system operator of a network configures link costs for each interface in a router, by using CLI)
• In order for routers to exchange OSPF messages with each other, IP addresses for identifying routers (OSPF RID (Router ID)) are required. For the purpose, loopback addresses are usually used. IP addresses of physical links can also be used, but loopback addresses are preferred for the following reasons:
  • If a physical link IP address is used: The IP address disappears when the physical link is down. That is, an OSPF router that uses this address as OSPF RID can no longer communicate with its connected neighbor OSPFs.
  • If a loopback IP address is used: The IP address remains valid even when the physical link is down because loopback addresses are logical addresses that are not bound to a certain physical link.

 

Routers flooding their link IP addresses to other routers

 

1. R1 floods its link information to neighbor OSPF routers

 

 

R1 floods its link (interface) address and loopback address to R2 and R3, through Router-LSA message [t=1]. Upon receiving the message, R2 and R3 flood it through all the links except the one that the message was sent through, having both R4 and R5 receive the same Router-LSA message twice [t=2]. At this time, if the two messages have the same LSA identifiers, {LS Type, Link State ID, Advertising Router}, the one that arrived later is discarded.

• LS Type: Identifies Router-LSA (1), Network-LSA (2), External-LSA (5), etc.
• Link State ID: Same as Advertising Router value (10.1.0.1) in case of Router-LSA
• Advertising Router: ID of the router that originated generated Router-LSA (OSPF RID, i.e. loopback address = 10.1.0.1) 

• R1's loopback address (10.1.0.1/32)
• R1's link address 1.1.1.1 is connected to R2 (10.1.0.2)
• R1's link address 1.1.2.1 is connected to R3 (10.1.0.3)

 

 

Routers building OSPF Topologies

 

 

As a result of exchanging Router-LSA messages with each other, all the five OSPF routers end up having the same LSDB, and thus the same OSPF topology map built based on the LSDB as seen above.

 

 

Calculating shortest path tree

 

 

Using the OSPF topology generated above, each router calculates the shortest path tree, i.e. the shortest path to their respective destinations. The shortest path information is installed at the Routing Information Base (RIB) and Forwarding Information Base (FIB) of each router (Note: RIB is a routing table managed in the control plane where routing protocols like OSPF are used while FIB is a routing table located in the data plane where wire-speed packet forwarding by packet processors takes place).