FORK: A novel two-pronged strategy for an agent-based intrusion detection scheme in ad-hoc networks

In this paper, we introduce FORK, a novel two-pronged strategy to an agent-based intrusion detection system for ad-hoc networks. We follow two different but complementary approaches for intrusion detection in our proposed scheme. We perform intrusion detection for power-aware ad-hoc networks. We introduce a novel power and reputation-based auctioning scheme for distributing agent-tasks in the network. Nodes compete for, and win auctions for performing the tasks based on a competitive power-efficient mechanism that permits collaboration between nodes. The chosen nodes perform the intrusion detection using our proposed anomaly detection algorithm that is modeled on popular evolutionary algorithms techniques. We evaluate our system both in terms of the task allocation algorithm as well as results of actual intrusion detection performed in some session log files. The outcome is promising and offers scope for some interesting additional research.