Compromise-Resilient Anti-Jamming Communication in Wireless Sensor Networks

Open Access
Jiang, Xuan
Graduate Program:
Computer Science and Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
Committee Members:
  • Guohong Cao, Thesis Advisor
  • Sencun Zhu, Thesis Advisor
  • Jose Antonio Ventura, Thesis Advisor
  • Raj Acharya, Thesis Advisor
  • Jamming
  • Compromise-Resilient
  • Sensor Networks
Jamming is a kind of Denial-of-Service (DoS) attack in which an adversary purposefully emits radio frequency signals to corrupt the wireless transmissions among normal nodes. Some research has been conducted on countering jamming attacks for wireless networks. Few works consider jamming attacks launched by insiders, where an attacker first compromises some legitimate sensor nodes to acquire the common cryptographic information of the sensor network and then jams the network through those compromised nodes. In this paper, we first retrospect research works for wireless networks from its feasibility to countermeasures. Then, we address the insider jamming problem in wireless sensor networks. In our proposed solutions, the physical communication channel of a sensor network is determined by the group key shared by all the sensor nodes. When insider jamming happens, the network will generate a new group key to be shared only by the non-compromised nodes. After that, the insider jammers are revoked and will not be able to predict the future communication channels used by the non-compromised nodes. Specifically, we propose two compromise-resilient anti-jamming schemes: the split-pairing scheme which deals with a single insider jammer, and the key-tree-based scheme which copes with multiple colluding insider jammers. We implement and evaluate the proposed solutions using Mica2 Motes. Experimental results show that our solutions have low recovery latency and low communication overhead, and hence they are suitable for resource constrained sensor networks.