Intelligent Behavioral Fingerprinting - From Theory to Practice

Tutorial session related to my PhD Thesis at CNSM 2021

Link

Alberto Huertas¹, Pedro M. Sánchez², Muriel Franco¹, Gérôme Bovet³, Gregorio Martínez², Burkhard Stiller¹

¹ Communication Systems Group CSG, Department of Informatics IfI, University of Zurich UZH, Switzerland ² Department of Information and Communications Engineering, University of Murcia, Spain ³ Cyber-Defence Campus within armasuisse Science & Technology, Thun, Switzerland

Previsions for 2025 estimate nearly 64 billion Internet-of-Things (IoT) devices connected via the Internet to diverse application scenarios. These scenarios show particularities in terms of devices, communications, data, and services, which increase the complexity of achieving one of their common challenges: to optimize the efficiency of their services. Today, a thriving challenge in behavior data science lies in creating behavior patterns (fingerprints) of devices and networks to optimize their performance and detect potential concerns, especially cyberthreats at early stages.

This tutorial provides an overview of tools and techniques toward behavioral fingerprinting and how these can be used to optimize the efficiency of heterogeneous application scenarios such as cybersecurity, or device authentication. Also, this tutorial explains well-known techniques used to create and evaluate fingerprints, paying particular attention to recent and promising Machine Learning (ML) and Deep Learning (DL) approaches. Further, this is discussed in detail within a set of real use cases, where behavioral fingerprinting is applicable to solve cybersecurity concerns of IoT-based application scenarios. Finally, a selected set of lessons learned within this dedicated area of security management is presented and future trends of behavioral fingerprinting are outlined.