A Novel Cross-layer Based Modified Support Vector Machine (E-svm) for Detecting Blackhole and Wormhole Attack in Wireless Ad-hoc Networks

Abstract

Abstract Wireless Ad-hoc Network do not require existing infrastructure .The devices generally that can exist are an access point that connects the end computing devices. Existing approaches suffer from challenges like low detection accuracy, low deviation structure and has more processing time. The paper provides a cross-layer approach by using physical and MAC layer knowledge of the wireless medium that is shared with the higher layer for handling the wormhole and black hole attack. The wormhole attack using tunneling whereas the black hole attack impersonates the source by changing the network traffic. The cross-layer capabilities are provided with three layers Network layer, MAC layer and physical layer and are independent of network protocols. The channel interference is handled by the physical layer, process handling by Network layer. Mac layer works with Network and physical layer by taking care of the Bandwidth information, Number of failed transmissions. The performance estimation for Time interval is in seconds. The proposed E-SVM Algorithm outperforms the SVM algorithm in terms of average consumed energy, average remaining energy, and packets received, packet delivery ratio, Delay, Jitter, Throughput, Normalized Overhead, Dropping ratio, Goodput. To do that E-SVM proposed and simulated in NS2.32 software and the results are demonstrated in terms of various qualities of service parameters and performance is evaluated.