In the recent years, wireless technology has enjoyed a tremendous rise in popularity and usage. One such field concerns sensor networks where the sensor nodes do not rely on any fixed network infrastructure. Many applications have been proposed here, such as military applications, industrial applications, and environmental applications. As sensor networks move closer to mission-critical deployments, security becomes a major concern. However, due to the resource-constraint, susceptibility to physical capture, and wireless nature of sensor networks, security is very difficult to achieve. Existing works in this field mainly rely on cryptographical schemes, however, they are not enough and sometimes not efficient. For example, most secure routing protocols proposed suffer from certain well-known attacks, e.g. compromised nodes attacks and denial-of-service attacks, and are not flexible in adjusting the tradeoff between network performance and security cost.
The thesis studies on the required and appropriate security services for wireless sensor networks. We develop PLUS_F-a Parameterized Localized trUst management based Security Framework for sensor networks, which is a novel approach from system architecture view. It parameterizes characteristics of application domain, status of network, and nodes' behavior; explores a distributed trust model, enabling recommendation-based trust and trustbased recommendation, to build reasonable trust relationship among network entities; and manages local knowledge to take countermeasures to minimize the effectiveness of an attack and keep the performance of the network within acceptable limits.
Then, we apply PLUS_F to design PLUS_R-a Parameterized Localized trUst-based Secure Routing for sensor networks. It parameterizes characteristics from routing perspective; lies in PLUS_F to enable trustworthy relationship establishment; and provides a secure routing using the available nodes' behavior evidence aiming to minimize network burden and maximize security functions.
Finally we analyze the network condition under various attack models by simulation, which can be used to choose more accurate and efficient system parameters; next, we estimate the extra cost caused due to the trust management; and finally we show that our secure routing which is based on our security framework can provide more secure routing paths with little increased packets transmission.