Design a framework for identification and confrontation of denial of service attack in wireless sensor network

Abstract

Wireless Sensor Networks (WSNs) has tiny sensor nodes to newlinemonitor the surrounding environment. It transmits this aggregated data to the newlinenearby Base Station for further processing. This sensitive data should be newlineprotected against various active and passive attacks. Denial of Service (DoS) newlineis an active attack, which intends to block the service of a network by creating newlineunwanted traffic resulting in draining the resources of the network. Due to newlineresource constraints and limitations, it is difficult for the developers to design newlinesecurity mechanisms for such networks. Though lot of authentication newlineprotocols have been implemented to comprehend this attack, they are not newlinereliable as they have no sufficient provision to predict this attack and protect newlinedata synchronization among participants. newlineTo address the above limitations, in this thesis two novel Elliptical newlineCurve Cryptography (ECC) based authentication protocols are proposed. In newlinethe first protocol, a novel proficient and DoS resisting user authentication newlinemethod is established by incorporating DoS identification and alleviation newlineissues. This protocol has the following phases: (i) User Registration newline(ii) Remedy (iii) Attack prediction and (iv) Reloading. Moreover, this newlinedeveloped approach chiefly considers the attack prediction phase, in which newlineNeural Network (NN) is incorporated to detect DoS attack. As a part of this, newlineNN is optimized by the selection of optimized weights. For training the NN, a newlinenovel enhanced optimization scheme called Fitness Indexed Whale newlineOptimization Algorithm (FI-WOA) model has been introduced. The newlineperformance of the adapted optimization scheme is examined over the newlineexisting techniques using various measures. newline