Policy-driven Physical Attacks in Sensor Networks: Modeling and Measurement


Sensor nodes being small in size and distributively deployed, are vulnerable to Physical Attacks that attempt to physically destroy sensors in the sensor network. Generally speaking, physical attacks in sensor networks can be classified into two types: Blind Physical Attacks and Search-based Physical Attacks. In blind attacks, sensors are destroyed using brute-force approaches (like bombs/grenades etc.). The advantage here is the rapidness in destroying sensors. The downside however, is the fact that the deployment field also suffers significant casualties. If the attacker wishes to preserve the deployment field, the attacker will conduct search-based attacks by searching for sensors in the field and destroying only the sensors. While this preserves the deployment field, the attack process is slow. In this paper, we present Policy-driven Physical Attacks, where the bias between the twin objectives of the attacker (rapidly destroying sensors, and preserving the deployment field) is modeled as a policy for the attacker. In policy-driven physical attacks, the attacker walks through the sensor network deployment field using signal detecting equipment to locate active sensors. Depending on the attacker's policy, the attacker takes different actions during the attack process. Based on detailed performance measurement, we observe that the policy has impacts on the network performance and destruction in the deployment field, demonstrating that the attacker can achieve desired bias in its objectives under policydriven physical attacks.


Computer Science

Keywords and Phrases

Measurement; Modeling; Physical Attacks; Policy-Driven; Sensor Networks

Document Type

Article - Conference proceedings

Document Version


File Type





© 2006 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.