Data dissemination and routing in a gossip based mobile environment

Abstract

In this paper we study the information spreading behavior of selected, mobility models combined with epidemic algorithms and the dynamic routing efficiency of gossip based routing algorithms on ad-hoc networks. Epidemic algorithms follow a paradigm of nature where simple rules dictate the spread of information by having a local view of the environment. In our work, three models of mobility, Random Waypoint, Gauss Markov and Clustering Motion have been combined with the SIR epidemic model. Behavior of the SIR model in a mobile environment is based on susceptible, infectious and recovered population of the mobile network. It can be seen that there is a significant drop in message overhead in the route protocol when combined with the gossip protocol and such reduction in message overhead is independent of the mobility model. However, there is an accompanied convergence time in create, which is dependent on the mobility model. I. INTRODUCTION Mobile network nodes operate under limited computational, communication and energy resources. As a node moves over the network it can be in different locations over a time period. To spread messages over the network we can use gossip algorithms with mobility models. The meaning of Gossip [4] is, sharing and spreading data among number of nodes through neighbor nodes in the network. There are several mobility models proposed by other researchers. Brownian model [2], Random Waypoint model [6], Multicast model [3] and Gauss Markov model are some models they have introduced. Gossip based algorithms for information spreading has recently received important attention for sensor and ad hoc network applications because of their simplicity and robustness. Common problem of many protocols is the wastage of energy in passing redundant information over the network.When the gossip model of communication is combined with a mobility model, it has different probabilities for periodic replica updates under different mobility models. These probabilistic values will be varying if nodes move as clusters. It should be discussed each and every state of the network and give a probabilistic model to clarify the above problems in gossips on mobility models.Ad hoc On-Demand Distance Vector (AODV) [14] is a method of routing messages between mobile computers. It allows mobile computers, or nodes, to pass messages through their neighbors to nodes with which they cannot directly communicate. AODV does this by discovering the routes along which messages that can be passed. AODV makes sure these routes do not contain loops and tries to find the shortest route possible.