Data-driven approach for automatic telephony threat analysis and campaign detection

The growth of the telephone network and the availability of Voice over Internet Protocol (VoIP) have both contributed to the availability of a flexible and easy to use artifact for users, but also to a significant increase in cyberecriminal activity. These criminals use emergent technologies to conduct illegal and suspicious activities. For instance, they use VoIP's flexibility to abuse and scam victims. According to (F. I. F.. N. D. N. C. R. D. Book, Available at: https://www. ftc. gov/news-events/press-releases/2016/12/ftc-issues- fy-2016-national-do-not-call-registry-data-book, accessed on: 27 August 2017), US government revealed receiving more than 5.3 million telephony abuse complaints in 2016. Based on this report, more than 226 million phone numbers were registered on the Do Not Call Registry list as not to receive tele-marketing calls. For instance, they use VoIP's flexibility to abuse and scam victims. A lot of interest has been expressed into the analysis and assessment of telephony cyber-threats. A better understanding of these types of abuse is required in order to detect, mitigate, and attribute these attacks. The purpose of this research work is to generate relevant and timely telephony abuse intelligence that can support the mitigation and/or the investigation of such activities. To achieve this objective, we present, in this paper, the design and implementation of a Telephony Abuse Intelligence Framework (TAINT) that automatically aggregates, analyzes and reports on telephony abuse activities. We deploy our framework on a large dataset of telephony complaints, spanning over seven years, to provide in-depth insights and intelligence about emerging telephony threats. The framework presented in this paper is of a paramount importance when it comes to the mitigation, the prevention and the attribution of telephony abuse incidents. We analyze the data and report on the complaint distribution, the used numbers and the spoofed callers' identifiers. In addition, we identify and geo-locate the sources of the phone calls, and further investigate the underlying telephony threats. Moreover, we quantify the similarity between reported phone numbers to unveil potential groups that are behind specific telephony abuse activities that are actually launched as telephony abuse campaigns. (C) 2018 The Author(s). Published by Elsevier Ltd on behalf of DFRWS.