Exploring techniques for constraining outputs of network events to guarantee users’ privacy

Details

Supervisors

Barbette, Tom ; Legay, Axel

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] en cybersécurité, à finalité spécialisée: conception et analyse de systèmes

Abstract

In an increasingly connected world, it is becoming more and more difficult to ensure one’s privacy. Data is being collected at all times, whether it be for statistical studies, recommendation engines, or malicious purposes. Most of us recognize the need for privacy as a fundamental human right but in today’s world, access to technology and information is equally important. In order to maintain our access to information through the internet in a secure and efficient way, network administrators have the need to monitor their networks for faults and security events. Such monitoring of network activity can obviously lead to breaches in the users’ privacy. Previous works have shown the difficulty of anonymizing network data for its release. Thanks to the Retina framework, we can think of network privacy in a different manner by attempting to render analyses of the network data private instead of relying on faulty techniques for anonymizing packet traces. In this thesis, we discuss the applicability of different methods to protect the users of a network from privacy breaches. We compare the traditional methods of deidentification with more recent techniques, and we propose different ideas as a basis to explore in future works on the subject.