Novel assessment metric and countermeasures for traffic attack threats in wireless sensor networks

In wireless sensor networks, all data packets are routed from the individual sensor nodes towards an in-situ base-station (BS). Such traffic pattern makes the BS vulnerable to adversary's attack. Basically, an adversary would intercept the ongoing transmissions and localize their sources. Then by employing traffic analysis techniques, an adversary would correlate the intercepted transmissions to uncover the data path which may lead to the location of the BS. Evidence theory is a well-known scheme that an adversary might use for traffic analysis. However, prior work considered only intercepted transmissions as evidences in the correlation process without factoring in the time of interception. In this paper, we argue that time-based correlation increases the accuracy of the traffic analysis and makes contemporary countermeasures ineffective. A novel technique is proposed to counter the time correlation and boost the anonymity of the BS. The technique imposes buffering delay at each relaying node on the data route in order to disturb the time correlation among consecutive transmissions. Our technique is validated through simulation.