User:Baidufriend/sandbox

CBRP: Cluster Based Routing Protocol

Summary:

Cluster Based Routing Protocol (CBRP) is a routing protocol designed for use in mobile ad hoc networks. The protocol divides the nodes of the ad hoc network into a number of overlapping or disjoint 2-hop-diameter clusters in a distributed manner. A cluster head is elected for each cluster to maintain cluster membership information. Inter-cluster routes are discovered dynamically using the cluster membership information kept at each cluster head. By clustering nodes into groups, the protocol efficiently minimizes the flooding traffic during route discovery and speeds up this process as well. Furthermore, the protocol takes into consideration the existence of uni-directional links and uses these links for both intra-cluster and inter-cluster routing.

Introduction:

There are several major difficulties for designing a routing protocol for MANET.

Firstly and most importantly, MANET has a dynamically changing topology due to the movement of mobile nodes which favors routing protocols that dynamically discover routes (e.g. Dynamic Source Routing, TORA , ABR etc.) over conventional distance vector routing protocols.

Secondly, the fact that MANET lacks any structure makes IP subnetting inefficient. However, routing protocols that are flat, i.e. have no hierarchy, might suffer from excessive overhead when scaled up.

Thirdly, links in mobile networks could be asymmetric at times. If a routing protocol relies only on bi-directional links, the size and connectivity of the network may be severely limited; in other words, a protocol that makes use of uni-directional links can significantly reduce network partitions and improve routing performance.

CBRP has the following features:

1. fully distributed operation.

2. less flooding traffic during the dynamic route discovery process.

3. explicit exploitation of uni-directional links that would otherwise be unused.

4. broken routes could be repaired locally without rediscovery.

5. sub-optimal routes could be shortened as they are used.

The operations of CBRP are entirely distributed. The major components are: Cluster Formation, Adjacent Cluster Discovery and Routing.

Cluster Formation:

The goal of Cluster Formation is to impose some kind of structure or hierarchy in the otherwise completely disorganized ad hoc network. The algorithm is a variation of the simple "lowest ID" clustering algorithm in which the node with a lowest ID among its neighbors is elected as the Cluster Head.

Adjacent Cluster Discovery:

Cluster X and cluster Y are said to be bi-directionally linked, if any node in cluster X is bi-directionally linked to another node in cluster Y, or if there is a pair of opposite uni-directional links between any 2 nodes in cluster X and cluster Y respectively.

Cluster X is said to be uni-directionally linked to cluster Y if they are not bi-directionally linked and if there exists some node in cluster X that is uni-directionally linked to some node in cluster Y. X is called Y's upstream uni-directionally linked adjacent cluster, and vice versa.

The goal of Adjacent Cluster Discovery is for a cluster to discover all its bi-directionally linked adjacent clusters. For this purpose, each node keeps a Cluster Adjacency Table (CAT) that records information about all its neighboring cluster heads.

Routing:

Routing in CBRP is based on source routing. It can be viewed as consisting of 2 phases: route discovery and the actual packets routing. Cluster structure is exploited to minimize the flooding traffic during route discovery phase. Furthermore, certain uni-directional links are discovered and used, thus increasing the network connectivity.