Ubiquitous wireless sensor networks are networks consisting of several small computers operational with sensors to discover events like human movement with infrared sensors or resolve the existing state of certain parameters like temperature. These sensor nodes are assembled with a radio to commune with each other and to broadcast data to a central computer where this data can be analyzed.
These new sensing computing machines come with new challenges. To eliminate the limitations imposed by wires, sensors possess limited energy sources and correspond with neighboring nodes using wireless networks. Due to critical nature of applications, securing data generated in sensor field by sensors is important. While data security is necessary, the functionality cost in terms of routing and energy needs to be assessed carefully.
Limited bandwidth and memory pull the trigger to make more sophisticated and light weight cryptography models. Moreover, deployment of tiny devices like sensor nodes in hostile environment, make it easier for intruder to budge in, plant attack node or control over the legitimate node to launch an attack. In this thesis, we have analyzed the basic threat model in ubiquitous wireless sensor network (USWN), and presented a quad tree based network portioning scheme, which ensure data accuracy at primary level. Each sensor node maintains neighboring table which stores the node information, our neighbor voting based defense solution helps to authenticate clear text broadcast of control packets. Once we have identified malicious node, our localization scheme can highlight the vulnerable node locality, even most of anchors are compromised. Localization scheme also helped us to strengthen defense against Sybil attack and wormhole attack. We analyze security of proposed scheme against Hello flood, data authentication attacks, node cloning attack and routing attacks.
In the existing scenarios, where diverse security services are independent from each other, each service adds an additive communication overhead. By incorporating energy-concentrated security services that are regularly required in sensor network applications, we considerably decrease the energy.