Study on Cross Layer Selfish Behavior and Detection Mechanism in Wireless Ad Hoc Networks
|Course||Applied Computer Technology|
|Keywords||Wireless Ad Hoc Networks IEEE 802.11 Selfish Behavior Cross Layer Detection|
Due to its distributed topology and organization of network components, together with the characteristic of fast and flexible establishment, wireless Ad Hoc networks are the most pervasively used communication environment. Medium Access Control (MAC) layer protocol is now employing a random access mechanism for contention on wireless channel. Nevertheless, with the development of network technologies and the restrictions to network resources, wireless stations may start to be selfish and violate the original protocol in the purpose of making more benefits for themselves. In recent years, the problem of diversified misbehaviors of selfish stations is turning into a later-model issue that jeopardizes network security and credibility, which has drawn increasing attention from researchers both home and abroad.Selfish stations may apply various strategies to increase their own share to the wireless channel in the form of maximizing their throughput or minimizing their delay, one of which is giving MAC layer a small back off value each time when they enter the binary exponential back off process. What’s more, some stations may do this in a cross layer way, with information from network layer. This dissertation analyzes the performance of selfish policies in both MAC layer and cross layer environment based on simulation results in Qualnet 4.0 to measure Average Medium Access Delay (AMAD) in MAC layer. Concluding from the above analysis, the AMAD of selfish stations decreases when they reduce the value of contention window, while that of well-behaved stations increases. As the number of selfish stations expands, the AMAD of both selfish and normal stations will increase.To point against such selfish behavior of wireless station, we develop a statistic-method based cross layer detection algorithm. During the detection period, each wireless station collects the Medium Access Delay series from its neighbors under promiscuous mode. It then filters the values with cross layer information, and finally compares the calculated average value of delays to the threshold value and determines whether other stations are selfish according to the comparisons. We also improve this algorithm by attaching supplementary reputation evaluation and penalty mechanisms to push stations to revise their behaviors and work in collaboration with others again. Simulation results show that the proposed algorithm has high detection veracity and low false alarm rate of the algorithm, and the reputation-based mechanism can improve cooperation between selfish stations and others.