Event detection and correlation for network environments

Autonomic communication has the aim of supportingfast-evolving network technologies and services, andof reducing the burden in managing complex and dynamicnetwork environments. Networks with desirable self-* propertiesshould be more adaptable to changing conditions and wouldenable greater flexibility and functional scalability. A necessarycondition for realising these benefits is a heightened level ofnetwork awareness; this requires not merely the capacity tomonitor the system and network state, but also the abilityto characterise the operational environment and its dynamicshifts. As argued in the literature, patterns of change can bedetected through cross-correlation of monitored or sensory inputsexpressed in events. The profiling of the temporal orderingand other relationships encoded in these patterns can providecontextual information suitable for reasoning and adaptationtasks. In this paper, we present the design framework andinitial evaluation of an Information Sensing system that aimsto enable awareness through an integrated event detectioncorrelationmechanism. In the context of autonomic networks,it offers a more lightweight solution than traditional activedatabase-oriented event systems, and it has better performancethan log-post analysis processing; its design also provides for adistributed detection facility.