To evaluate the reliability of WSNs (Wireless Sensor Networks) in the malware spread scenario

the traditional epidemic model SEIR was extended by introducing a state "Dead" and the dynamic change process of sensor node states was exactly described by a Markov chain (MC). Using the stochastic game

a malware spread detection model for WSNs was proposed to predict the probability of malware adopting the spread behavior. The prediction results attained were integrated into the MC's state transition matrix in order to relate the malware intention to the MC randomness. An equation to compute a sensor node's reliability in the malware spread scenario was constructed via the MC

and reliability evaluation methods for WSNs with star or cluster topology were given. Experiments have shown the influence of malware's spread on sensor nodes' reliability and illustrated that the proposed method can efficiently evaluate the WSNs reliability.