Optimal wireless sensor network secure routing for pervasive computing

Abstract

Wireless Sensor Networks (WSNs) contain small low power nodes (sensors) that can be deployed and embedded in a certain area and have the ability to provide monitoring and interacting with the physical environment. Deploying and embedding such nodes in the physical environment provides the visibility of pervasive computing. In pervasive computing software applications are aware of environmental conditions pertinent to their operation, and adjust their functionality as these conditions change. However, security, processing and power limits of WSNs are still the main challenges in this area. The literature review has revealed that while many solutions and contributions have proposed solutions to these problems but still some gaps exist. First, there has been limited research on addressing the factors of power consumption which reduces network lifetime; in this research the power efficiency of WSNs is studied. This research addresses three routing parameters that affect the network lifetime such as the initial power of the nodes, the residual power in the nodes, and routing period. Simulation results of effects of these parameters individually and collectively presented in different network sizes. Addressing the effects of these factors on the network lifetime is the main contribution to improve network lifetime through consideration of power consumption factors during routing. Second, on the issue of security, many studies have proposed trust and reputation mechanisms to increase reliability and privacy of networks. However, the literature has shown a lack of research that compares existing trust and reputation models especially in terms of energy efficiency. This study presents a comparison between several proposed trust and reputation models in terms of average path length, average accuracy, and energy consumption rate. Then, this study proposes a reputation and trust mechanism optimized for security strength. During simulation of the model two security threats (oscillating and collusion) were applied in order to measure the accuracy, scalability, trustworthiness and energy consumption. As a result, the effects of collusion and oscillating are minimized and energy consumption for dynamic networks reduced. Also, simulation results show that the proposed model remains resilient to low or high percentages of pernicious servers when the percentage of client sensors is greater than or equal to 60%. This result is quite promising; and shows that energy consumption generally is low, especially for dynamic networks. Finally, energy consumption is the main challenge that shortens network lifetime and minimizes network efficiency. This study also proposed a hybrid secure and energy aware routing approach that considers the trustworthiness, energy level for sensors, efficiency and maximizing network lifetime. Performance and efficiency are measured through simulation results, analysis and comparison with other well-known mechanisms.