Improving SS7 Security Using Machine Learning Techniques
Master thesis
Permanent lenke
http://hdl.handle.net/11250/2403095Utgivelsesdato
2016-08-31Metadata
Vis full innførselSamlinger
Sammendrag
The Signaling System No. 7 is the nervous system of telecommunication networks based on 2G and 3G technologies. Previously confined in a walled garden, SS7 has become more exposed due to increased liberalization of the market in conjunction with the industry switching to IP technology. In the walled garden of trusted operators, security have received minimal attention. SS7 has become more vulnerable in the recent years, with attackers exploiting network communications to track subscribers, intercept calls, perform denial of services, and commit fraud. This master thesis is a part of the effort to reduce the vulnerabilities contained in the old, yet crucial protocols that the telecommunication operators cannot function without. Subscribers, operators, and national governments are dependent on one of societies critical infrastructures, it needs to be adequately protected. In this thesis, a detailed overview of SS7 threats and vulnerabilities is presented. In an effort to mitigate these attacks, open source technology has been used to simulate network traffic. This generated traffic were used to analyse and detect attacks against SS7 in an effort to propose detection mechanisms. Machine learning, big data, and anomaly detection techniques have been used as tools in order to propose an improved online protection system for SS7 networks. The results achieved in this master’s thesis have been submitted in the form of a paper to the International Conference on IT Convergence and Security 2016, Appendix A presents the submitted paper in its current form.