Give


Graduate Exam Abstract


Divyanka Bose

M.S. Final

July 30, 2015, 1:00 pm - 3:00 pm

ECE Conference Room

Security of Virtual Coordinate based Wireless Sensor Networks


Abstract: Wireless Sensor Networks (WSNs) perform critical functions in many applications such as military surveillance, rescue operations, detection of fires and heath care monitoring. In these applications, nodes in the network carry critical and sensitive data. Thus, WSNs are prone to various kinds of attacks that target different protocols and layers of the network. Also, most of the WSNs are placed remotely that makes it difficult to implement security measures after deployment. Thus, security of WSNs needs to be considered at the initial stage of system design. In many applications, the nodes are deployed randomly, and thus are unpredictable in terms of physical network topology. Virtual Coordinate (VC) based WSNs possess significant advantages over Geographical Coordinate based WSNs. This is because VCs negate the need for physical localization of nodes, which require costly techniques like GPS. The VCs of the nodes in the network are very important for basic functionalities such as routing and self-organization. However, security of VCs has not been extensively researched even though routing algorithms rely on the correctness of the VCs for proper functioning. VC based WSNs are susceptible to attacks resulting from malicious modification of VCs of individual nodes. While the impact of some such attacks is localized, others such as coordinate deflation and wormholes (tunneling) can cause severe disruptions. This thesis proposes techniques for the detection and mitigation of attacks, which are aimed at the Virtual Coordinate (VC) based WSNs. We propose a novel approach where coordinate attacks are identified by detecting changes in shape of the network, extracted using topology maps. A comprehensive solution for detection of coordinate-based attacks on VC systems is presented that combines Beta Reputation System and a reputation based routing scheme. Latter ensures safe communication that bypasses malicious nodes during detection process. The coordinate deflation and wormhole attacks are discussed and the effect and intensity of these attacks are addressed. Two methods are proposed and compared for the detection of attacks. In the first method, the topology distortion is rated using clusters identifiable by existing VCs, thus requiring low computation and communication overhead. A measure of topology distortion is presented. The existence of a trusted base station is needed for this method. In the second method, the detection is distributed and removes the need for a base station/server. We compare the advantages and disadvantages of the two methods, and discuss the scenarios in which these algorithms maybe implemented. Simulation based evaluations demonstrate that both the schemes efficiently detects deflation and wormhole attacks. We choose a variety of dense networks with different topologies and deployment characteristics for evaluation. Networks with voids, representative of physical spaces with voids, are considered, as well as randomly deployed networks to ensure the correct operation and scalability of the algorithms. We show through simulations that the detection schemes can easily differentiate between the changes in the network due to node failures, e.g., caused by battery drain, from those due to an attack. Future sensor networks are predicted to be in the scale of millions of nodes. Thus, a need for security algorithms which can be scaled are highly desirable. We show in our simulations that the proposed detection schemes can be applied to networks of larger density successfully.

Adviser: Dr. Anura Jayasumana
Co-Adviser: N/A
Non-ECE Member: Dr. Christos Papadopoulos
Member 3: Dr. Sudeep Pasricha
Addional Members: N/A

Publications:
D. Bose and A. P. Jayasumana, "A Reputation-Based Method for Detection of Attacks in Virtual Coordinate Based Wireless Sensor Networks," In Proceedings of 40th Annual IEEE Conference on Local Computer Networks(LCN), Oct 2015


Program of Study:
ECE520
ECE513
MATH560
ECE514
CS457
ECE530
ECE658
ECE699