A Survey on Distance Vector Routing Protocols

In computer communication theory relating to packet-switched networks, a distance-vector routing protocol is one of the two major classes of routing protocols, the other major class being the link-state protocol. A distance-vector routing protocol uses the Distributed Bellman-Ford algorithm to calculate paths. [13] Compared to link-state routing protocols, distancevector routing protocols only requires the nodes to store next hop and distance information for routing destinations while in a link-state routing protocol, a node has to store all the information about the network topology. And when the metric of a link changes, distance-vector routing algorithms only require the associated node to inform its neighbors while a link state routing protocol requires the node to broadcast the information in the whole network. So typically distance-vector protocols have less computational complexity and message overhead.

However, distance-vector routing protocols also have several disadvantages. The most famous are the routing loop and count to infinity (CTI) problem. Because each router only has limited information about the network topology, routing loops might emerge and lead to CTI problem, greatly impede the efficiency of the protocol. In this survey we will see how several distance vector routing protocols have worked to alleviate or solve this problem. Distance vector routing protocols also suffers from security issues, because routing computation is done distributively, a malfunctioning or malicious node may severely affect the whole network. However, in recent years more secure distance vector routing protocols have been proposed [7].Another critical issue is the support for routing areas. In reality, large networks are typically divided into areas to accelerate routing, but distancevector routing protocols don’t support routing areas, so they are not suitable for really big networks.

In general, distance vector routing protocols are more suitable for small or median sized networks or when each node only have a limited storage or computing power. In reality, RIP and EIGRP are two very successful Interior Gateway Routing Protocols and major competitors for the link-state OSPF routing protocol. Now we’ll begin our introduction to RIP, AODV, EIGRP, RIP-MTI and Babel routing protocols. We’ll also give some of our thoughts on how to solve the CTI problem(See Section 7). In the end we summarize the results in Table 1.

The Routing Information Protocol is so far the most popular distance vector routing protocol, and perhaps the most popular interior routing protocol is the TCP/IP suite. RIP protocol’s popularity results from its simplicity, early adoption in a popular operating system (BSD) and early standardization in the RFC.

RIP protocol is the simplest form of a distance vector routing protocol. For example, the routing table is very simple and only employs very limited information-destination, hop count and next hop.

Its working are also easy to understand. On a regular basis, each router in the network sends out its routing table to its neighbors, informing them to which subnets it is connected and how far these subnets are (by measure of hop count). Once a router receives such a routing message, it updates its routing table. Say router B sends to router A claiming it has a route to C with hop count K, then A knows that by going through B, it can reach C with hop count K + 1. RIP protocol is essentially based on a distributed version of the famous Bellman-Ford shortest path algorithm. Assuming the protocol is executed in a synchronous fashion (that is, each round the routers receives the routing message, updates the routing table and sends out a new message, at the same time), then Bellman-Ford algorithm tells us the routing path will converge in no more than K rounds, where K is the diameter of the network.

The RIP protocol has several advantages. It is simple, and it is every efficient for small and simple networks, consuming little network bandwidth and little storage and computing power for the routers (which may be a battery powered small device).

The RIP protocol, however, also suffers from some inherent limitations. The most famous is perhaps the count to infinity problem. In Figure 1 say A,B and C are all connected in a network and suddenly C is disconnected because of a corrupted link. We would naturally want A and B to find this out immediately. Now note that B initially has a route to C with hop count 1. Now this route is invalid, but instead B would find that A has a route to C with hop count 2, and so it assume it has a route to C through A, with a hop count of 3; then, A would similarly update its route to C with a hop count of 4; and the Figure 1: a simple network topology that would cause count to infinity problem process goes on. This kind of routing loop causes the count to infinity problem, draining the network bandwidth, slowing down the routing path convergence and severely impeding

This content is AI-processed based on open access ArXiv data.