Introducing epidemic models for data survivability in unattended wireless sensor networks

One of the most relevant issues pertaining UWSN is to guarantee a certain level of information survivability, even in presence of a powerful attacker. In this paper, we provide a preliminary assessment of epidemic-domain inspired approaches to model the information survivability in UWSN. In particular, we show that epidemic models can be used to set up the parameters that allow the information to survive, once estimated the maximal compromising power of the attacker. Further, we point out that the mere application of these models is not always the right choice. Indeed, it comes out that these deterministic models are not enough accurate, and "unlikely" events can cause the loss of the datum. Finally, we provide some final comments, as well as promising research directions. © 2011 IEEE.