Mining Complex Event Patterns in Computer Networks

More and more ubiquitous and mobile computer networks are becoming available, which leads to a massive growth in the amount of traffic and according log messages. For handling and managing networks efficiently, sophisticated approaches for network management and analysis are necessary. In this paper, we show how to use temporal data mining in a declarative framework for analysing log files for computer networks. From a sequence of network management protocol messages, we derive temporal association rules, which state frequent dependencies between events. We also present methods for extendable and modular parsing of text messages and their analysis in log files based on XML.