A Privacy preserving Approach to Call Detail Records Analysis in VoIP Systems

Attacks on Voice-over-IP calls happen frequently. A specific type of these attacks are toll-fraud attacks. The prevention of these attacks depends on understanding the attack patterns. These can be derived from communication records. However, these records contain privacy relevant information of the call participants. These records are also protected by a number of laws and regulations. To make an analysis privacy compliant, relevant laws and regulations need to be considered. We propose a method for changing communication records in such a way that the forensic analysis in VoIP attacks is possible and the privacy of the call participants is preserved. We define privacy requirements for communication records from laws, regulations and concerns of call participants. We also present patterns of communication records based upon real world examples. We further show a framework for privacy attack identification and privacy data minimisation for a structured analysis of communication records. Moreover, an analysis pattern for toll-fraud attacks states which relations in the communication records have to survive the data minimisation.