부산대학교 도서관

Worms spreading directly between cell-phones over short-range radio (Bluetooth, WiFi, etc.) are increasing rapidly. Communication by these technologies is opportunistic and has very close relation with the social characteristics of the phone carriers. In this paper, we try to evaluate the impact of different characteristics on the spreading performance of worms. On the other hand, the behaviors of worms may have certain impact, too. For example, worms may make phones be completely dysfunctional and these phones can be seen as killed. We study the impact of the killing speed. Using the Markov model, we propose some theoretical models to evaluate the spreading performance in different cases. Simulation results show the accuracy of our models. Numerical results show that if users do not believe the data coming from others easily, the worms may bring less damage. Surprisingly, if the users are more willing to install the anti-virus software, the worms may bring bigger damage when the software becomes to be outdated with high probability. Though the worms can bring big damage on the network temporarily by killing phones rapidly, numerical results show that this behavior may decrease the total damage in the long time. Therefore, killing nodes more rapidly may be not optimal for worms.