How Routing Protocols Work

Every network routing protocol performs three basic functions:

Discovery: Identify other routers on the network.Route management: Keep track of the possible destinations (for network messages) along with some data describing the pathway of each.Path determination: Make dynamic decisions for where to send each network message.

A few routing protocols (called link-state protocols) enable a router to build and track a full map of all network links in a region while others (called distance-vector protocols) allow routers to work with less information about the network area.

The network protocols described below each allow computer routers to communicate with each other while forwarding traffic between networks. They are among the most popular protocols used.

RIP

Researchers developed Routing Information Protocol in the 1980s for use on small- or medium-sized internal networks that connected to the early internet. RIP is capable of routing messages across networks up to a maximum of 15 hops. RIP-enabled routers discover the network by first sending a message requesting router tables from neighboring devices. Neighbor routers running RIP respond by sending the full routing tables back to the requestor, whereupon the requestor follows an algorithm to merge these updates into its own table. At scheduled intervals, RIP routers then periodically send their router tables to their neighbors so that any changes can be propagated across the network.

OSPF

Open Shortest Path First was created to overcome some of the limitations of RIP, including:

15 hop count restriction.Inability to organize networks into a routing hierarchy, important for manageability and performance on large internal networks.Significant spikes of network traffic generated by repeatedly re-sending full router tables at scheduled intervals.

OSPF is an open public standard with widespread adoption across many industry vendors. OSPF-enabled routers discover the network by sending identification messages to each other followed by messages that capture specific routing items rather than the entire routing table. It is the only link-state routing protocol listed in this category.

EIGRP and IGRP

Cisco developed Internet Gateway Routing Protocol as another alternative to RIP. The newer Enhanced IGRP (EIGRP) made IGRP obsolete starting in the 1990s. EIGRP supports classless IP subnets and improves the efficiency of the routing algorithms compared to older IGRP. It does not support routing hierarchies, like RIP. Originally created as a proprietary protocol runnable only on Cisco family devices, EIGRP was designed with the goals of easier configuration and better performance than OSPF.

IS-IS

The Intermediate System to Intermediate System protocol functions similarly to OSPF. While OSPF became the popular choice, IS-IS remains in widespread use by service providers who have benefitted from the protocol being adaptable to their specialized environments. Unlike the other protocols in this category, IS-IS does not run over Internet Protocol (IP) and uses its own addressing scheme.

BGP and EGP

The Border Gateway Protocol is the internet standard External Gateway Protocol (EGP). BGP detects modifications to routing tables and selectively communicates those changes to other routers over TCP/IP. Internet providers commonly use BGP to join their networks together. Additionally, larger businesses sometimes use BGP to connect multiple internal networks. Professionals consider BGP the most challenging routing protocol to perfect due to its configuration complexity.