A Lightweight Authentication Protocol for IoT-Based Applications Using Reconfigurable Noisy PUFs
Abstract
Managing security threats is one of the most significant challenges when implementing Internet of Things (IoT) networks. Furthermore, most IoT applications, such as industrial IoT (IIoT), use resource-constrained devices that are also not physically protected. These facts pose severe difficulties in designing efficient security protocols. Some recent studies proposed lightweight authentication protocols based on physically unclonable functions (PUFs) to solve the mentioned problems. While PUF-based protocols are promising for security applications, there are practical challenges that must be addressed. These challenges include PUFs' noisy behavior and vulnerability to machine learning (ML) attacks, which have not been adequately considered in most current schemes. This article proposes an efficient and lightweight authentication protocol based on PUFs that addresses both issues and is, therefore, suitable for a wide range of IoT and IIoT applications. Formal security analysis (based on Burrows-Abadi-Needham logic) and informal analysis demonstrate that our proposed scheme appropriately provides security features. We also evaluate our protocol's computational complexity and communication overhead and demonstrate its computational efficiency and superior security features compared to similar schemes.