A Model for Global Anomaly Detection in a Grid Environment

Grids being concerned with sharing of diverse resources in a distributed environment induce security concerns. An event that occurs locally on a grid node may actually be part of a larger set of events that are collectively harzadous to the grid. Since a node only sees the local footprint of an event, it cannot know the contribution of this event at a global scale. As a result, several events that appear innocuous at any given node are ratified as safe, even though they may eventually contribute to a hazard at a global scale. To address such problems, we propose a model for anomaly detection in a grid environment by judiciously exchanging local profiles to capture global behavior. The key focus of this model is to detect anomalous behavior, which looks normal locally at any individual grid node, but when observed globally the anomalous behavior is apparent.