A Note on Statistically Detecting Tampered Type Attacks

Information integrity of communication data is crucial to e-business systems. An adversary may delete or insert packets into a normal communication data series in e-business systems to destroy information integrity to produce deserters (likely informal, this paper uses the term tamped type attacks to specifically describe those that attackers detect or insert packets into a normal data series). Consequently, a security issue in an e-business system is how to detect tamped type attacks. This short paper gives a method for statistically real-time detection of tampered type attacks. The method is based on power spectra of monitored traffic time series with finite length. The detection probability is derived.