Diversity for Detecting Routing Attacks in Multihop Networks

This paper presents a diversity model for reliable transmission in multihop networks under routing attacks. Traditionally, diversity is envisioned to be best achieved by utilizing multiple independent and disjoint paths. In this paper, we show that increasing the dependency/connectivity among transmission paths enables accurate detection and identification of malicious nodes. We define a diversity metric that reflects the degree of connectivity among routing paths. Then, we analyze the detection accuracy as a function of network diversity and communication overhead in a generalized network topology. Our results provide insights on network architecture and topology designs that should be tailored to achieve a good tradeoff between network reliability and transmission efficiency.