Self-similar traffic Analysis in Mobile Ad Hoc Network

Measurements of data traffic in real networks had shown character of self-similar which was different from traditional Markov hypothesis. Mobile ad hoc network also possessed this character in real network neglected by most researches which went against to apply theory to reality. This paper presents real limited G/D/1 queuing model to analyze queuing performance. The results showed that self-similar traffic undermines queuing performance and delay. This paper proposed a routing algorithm based on programming model for self-similar traffic, which exerted multi-path backup routing and reduced the total transmission power to weaken self-similar. The simulation results showed that the new algorithm can increase the throughput and improve packet delivery success ratio and delay is close to condition of Markov assume when nodes move at high speed