A fault localized scheme for false report filtering in sensor networks

Abstract

Sensor networks frequently deploy many tiny and inexpensive devices over large regions to detect events of interest. It can be easy to compromise sensors, enabling attackers to use the keys and other information stored at the sensors to inject false reports, forging fake events. Existing approaches do not localize the impact of such node compromises, so that compromises in one sensing region may compromise other parts of the system. In this paper, we propose two fault localized schemes for false report filtering. In our basic scheme, sensors signal events using one-way hash chains, which allows en-route nodes to verify the authenticity of received reports based on commitments of detecting sensors, but prevents them from forging events. We extend this basic scheme to a collaborative filtering scheme using commitment predistribution, making it more adaptable for mobile sensor networks and high-density sensor networks. Our scheme can also provide localized protection for areas that require special protection. Our security analysis shows that our schemes can offer stronger security protection than existing schemes, and are efficient.