Data-Driven Scalable Mechanisms and Architectures for Secure IoT Connectivity

The widespread adoption of Internet of Things (IoT) changed the way we live and work. It also introduces new challenges and risks that can impact individuals, organizations, and society at large. This paper introduces a novel approach to secure IoT connectivity using Named Data Networking (NDN), a paradigm well-suited for IoT due to its in-network caching and inherent security features. Recognizing the vulnerabilities of NDN to specific security threats such as Side-channel Timing Attacks (SCTA), Cache Pollution Attacks (CPA), and Interest Flooding attacks (IFA), we have embarked on a journey to enhance its security framework. Our approach encompasses the development of a comprehensive dataset for these attacks, the application of machine learning for accurate anomaly detection and classification, the introduction of a rate-limiting mechanism using Hidden Markov Models (HMMs) to combat IFAs, and the ongoing work on a cache replacement policy to effectively counter cache-based attacks.