As the computing paradigm is increasingly shifting from traditional client-server models towards edge computing and social computing, computer networks are becoming more and more complex and architecturally different from traditional infrastructure based IT networks. Sybil attacks, in which malicious actors insert themselves in a network and try to influence its behavior by compromising the networks’ trust underpinnings, pose major threats to these networks.  Traditional approaches of monitoring may not be appropriate specially in situations where privacy of the users of the networks is a serious concern. In this talk, we discuss our recent efforts towards addressing this problem. We model a complex network as a connected graph and develop techniques to identify its underlying structure. Several structural metrics are then used to identify nodes that play major roles on the stability of the network. Next, based on the topological properties of the graph, we develop similarity metrics for pairs of nodes that lead to exposing a community structure for the network. Using this community structure and a modified Random Walk algorithm, we show that it is possible to clearly identify nodes that behave significantly differently from typical nodes in the network.