The growing popularity of mobile networks makes them increasingly attractive to virus writers, and malicious code targeting mobile devices has already begun to appear. Unfortunately, standard techniques for modeling computer virus propagation cannot be applied to mobile settings. We describe why these models fail and introduce a new framework called probabilistic queuing which treats node mobility as a first-order concern. A network is modeled by multiple queues which emulate the skewed connectivity levels common in mobile environments. Each queue represents a separate epidemiological population, and as nodes shuttle between queues, they bring their infections with them. Simulations show that for realistic mobility parameters, our model is more accurate than the standard Kephart-White framework.
Search
Reviews
[Write a review of this article]
Find related articles from these CiteULike users
