Multi-Level Fusion to Improve Threat Pattern Recognition in Cyber Defense

Considering fast growth of internet and related network infrastructures, it is important to detect the intrusion and respond to it in a timely manner. Network intrusion can make vital information systems and communication networks inaccessible and imposes high cost of communication infrastructures. In order to gain high degrees of success in providing services, current and future generation of networking and internet technologies, require a set of tools to analyze the network and to detect the threats and intrusion in network. Due to main weakness in terms of high rate of false alarms and low accuracy of detection, by which cyber space detection and identification systems are opposed, fusion theory in decision level provides a new method for data analysis from multiple nodes in order to increase the possibility of intrusion detection through improving pattern recognition. This paper aims to present a novel method of fusion in decision level based on complex event processing and show how this method would be successful in exposing cyber threats for timely response.