A Model for Detecting “Global Footprint Anomalies” in a Grid Environment

摘要: Grid computing environments pose unique security concerns that are not generally relevant for conventional data management systems. An event appears as benign on a grid node, may actually be part of larger incident hazardous to the grid. Since node only sees local footprint an event, it cannot know contribution this at global scale. The focus work is detecting such kinds anomalous behaviors we call anomalies. In paper, propose two classes anomalies, and model detection class anomalies main challenge here detect behavior, which looks normal locally any individual but when observed globally, behavior apparent.