作者: R. Meena , S. Gopalakrishnan , A. S. Syed Navaz
DOI:
关键词: Data mining 、 Network element 、 Anomaly (natural sciences) 、 Computer science 、 Internet traffic 、 Anomaly detection
摘要: Introducing Internet traffic anomaly detection mechanism based on large deviations results for empirical measures. Using past traces we characterize network during various time-of-day intervals, assuming that it is anomaly-free. Throughout, compare the two approaches presenting their advantages and disadvantages to identify classify temporal anomalies. We also demonstrate how our framework can be used monitor from multiple elements in order both spatial validate techniques by analyzing real with time-stamped