Presence, Intervention, Insertion: Unifying Attack and Failure Models in Wireless Sensor Networks

As assumptions about adversaries critically influence the correctness and efficiency of protocols, they should be as precise as possible. We propose a general framework for attacker models in wirless sensor networks. The framework is modular since it allows to compose basic attacker models by choosing values from three orthogonal dimensions: presence, intervention, and insertion. All choices of basic attacker models form a lattice according to a well-defined weaker-than relation. Sets of such basic attacker models constitute general attacker models. In a sense, our framework unifies behavioral aspects of attacker models from the area of cryptography and fault-tolerance We show that our modeling framework can be used to formulate all relevant attacker models from the literature on sensor networks. We demonstrate the benefits of our framework by showing how it can be used to (1) compare protocols and establish relations between protocols and (2) to make attacker assumptions more precise and find errors in protocols.