Many post-quantum cryptographic schemes have been proposed in the literature, offering resistance to attacks launched by quantum computers. Due to its promising security properties, NTRU was selected as a candidate in the final round of the NIST competition on post-quantum cryptography. However, this method poses challenges for constrained devices due to its high computation and memory requirements. Motivated by the need to extend the lifetime of resource-constrained IoT devices while still being able to withstand a quantum attacker, in this chapter we propose a new NTRU-based collaborative encryption scheme for such devices. Our scheme preserves the confidentiality of sensitive information exchanged among constrained IoT devices deployed in an edge computing architecture. Our approach distributes the workload of the cryptographic operations among the edge nodes and IoT devices within the same network. This collaborative approach enables IoT devices to significantly reduce their computational cost while maintaining data confidentiality. Additionally, the proposed distribution of computing allows scalability and a better sustainability of IoT environments. Finally, we demonstrate through empirical evaluation and extensive simulation that our scheme is secure, improves energy saving and resource utilization in IoT networks, and is more efficient than the traditional centralized approach.

Trustworthiness is a critical element influencing every entity in fog computing architectures, which raises reliability and security issues for data exchanged between fog nodes and data sources. Malicious entities can introduce incorrect or manipulated information, requiring trust assessment and management. Maintaining trust throughout the fog computing architecture requires robust trust mechanisms, secure communication protocols, constant trust management and evaluation models. The integration of blockchain technology also provides a transparent and traceable environment, enabling the preservation of trust scores and promoting accountability. Indeed, we propose a new way of calculating event trustworthiness by exploiting trust factors such as Dempster-Shafer theory, temporal relevance and distance relevance. As a result, malicious entities can be discovered and trustworthy behavior can be promoted in the fog computing environment. Finally, we prove that our protocol is highly efficient, as demonstrated by extensive simulations and in-depth theoretical analysis that confirmed its security.

Federated identity management (FIM) is a promising approach that allows secure access to multiple security domains in heterogenous architecture such as fog computing. Keeping track of users connections in such contexts is however very challenging. Indeed, most of existing solutions do not provide connection auditing across multiple domain and thus cannot be adopted in federated identity protocols. Therefore, in this paper, we propose a novel connection tracking protocol for federated identity management systems in heterogeneous fog computing architecture. Our protocol creates a federation among multiple heterogeneous security domains and allows to keep track on users connections in a completely decentralized manner. As a result. Based on Blockchain Technology and robust cryptographic signature mechanisms, our protocol ensures an immutable global view of connections history all over the architecture. The Blockchain content is validated through a new distributed and secure consensus mechanism based on identity-based signatures. Allowing then a fast connection verification, recording and tracking, as shown in our performance evaluation.

Zero-day attack detection solutions aim to proactively identify unknown threats targeting valuable assets within a given system. While many Intrusion Detection System (IDS) solutions leverage learning techniques to build novel attack detection systems, they often focus on enhancing accuracy for specific attack types, overlooking the potential for multiple attack scenarios. Therefore, we introduce a novel framework for detecting zero-day attacks that evade current detection systems. Our framework enhances attack identification and qualification through a hybrid learning approach, where supervised learning ensures detection of known attacks and unsupervised learning. It encompasses intrusion detection phases from data collection to new attack class detection by identifying anomalies in real-time network flow data. Unsupervised learning, which involves grouping similar data points into clusters, establishes minimum distances within these clusters. This process triggers cluster division when certain thresholds are reached. Finally, an online supervised learning process validates our approach’s effectiveness in identifying anomalies associated with zero-day attack flows. This approach significantly reduces the False Detection Rate (FDR) without solely focusing on optimizing machine learning (ML) and deep learning (DL) algorithms hyper-parameters. We evaluated our framework on two datasets: one from a real industrial context at IBM and the NSL-KDD dataset. The results demonstrate our framework’s ability to detect anomalies in previously zero-day attack targets. On average, we identified 71 anomalous flows per target, achieving an overall average online learning accuracy of 98.4% for the IBM dataset and 96.6% for the NSL-KDD dataset, thereby validating the detection of these new attack scenarios.

We propose FLASH-RL, a framework utilizing Double Deep Q-Learning (DDQL) to address system and static heterogeneity in Federated Learning (FL). FLASH-RL introduces a new reputation-based utility function to evaluate client contributions based on their current and past performances. Additionally, an adapted DDQL algorithm is proposed to expedite the learning process. Experimental results on MNIST and CIFAR-10 datasets demonstrate that FLASH-RL strikes a balance between model performance and end-to-end latency, reducing latency by up to 24.83% compared to FedAVG and 24.67% compared to FAVOR. It also reduces training rounds by up to 60.44% compared to FedAVG and 76% compared to FAVOR. Similar improvements are observed on the MobiAct Dataset for fall detection, underscoring the real-world applicability of our approach.

Industrial IoT (IIoT) networks must provide reliability, determinism and security. In terms of security, ensuring an efficient authentication is still a challenging task. Most of the authentication approaches adopted in IIoT wireless communication protocols rely on the existence of a pre-shared key (PSK) between each joining node and the central authority of the network. How to share the PSK is however not specified in their standards. In this paper, we propose a new zero-touch mutual authentication and key establishment protocol for IIoT. In our protocol, the network coordinator authenticates a new joining node using certificates. Then, the joining node authenticates the network coordinator through a novel consensus achieved among nodes that are already in the network. Our novel consensus is based on Shamir secret sharing, and allows each new node to build its trust based on the knowledge of a group of nodes that are already in the network. Finally, our protocol allows to securely establish a common key between new joining nodes and the network coordinator over a public channel. In addition to the theoretical aspects, we evaluate the performance of our protocol under two attack scenarios where 33% of the nodes in the network are malicious. We show in all cases that we efficiently ensure authentication with high success probabilities, by establishing a consensus including a limited number of nodes in the network.

Fog computing is a promising computing paradigm that provides computing services close to end-users at edge of the network. Therefore, it supports large-scale, geographically distributed and latency-sensitive applications. However, due to its un-trusty nature, numerous security challenges must be overcome. The integrity verification of the data in fog computing context is one of the main challenges to tackle. Indeed, most of existing solutions consider a centralized data storage context, and usually rely on a trusted third party to check data integrity. However, these approaches are not suitable for fog computing architecture, since the data is stored in a dynamic and a completely distributed manner. Motivated by these challenges, we propose in this paper, a new efficient public verification scheme that protects the integrity of the data in fog computing architecture. Our scheme secures data integrity and authenticity based on the short integer solution SIS problem and the identity-based signatures. In addition, it allows to efficiently verify the integrity of data, even when it is separately shared on multiple servers. This verification can be performed by any end-user on the architecture, and without relying on any trusted third party. Finally, we show through extensive simulation that our solution is highly effective and outperforms existing solutions.

The Internet of Things (IoT) is a network that integrates a variety of heterogeneous nodes, such as connected devices (sensors, robots, and smart phones ...), connected cars, smart homes, etc. These smart objects communicate and collaborate in distributed and dynamic environments that are facing several security challenges. Trust management is one of the most important challenges in IoT. Existing trust management solutions do not meet the new requirements of IoT such as heterogeneity, mobility, and scalability. In this article, we propose a hierarchical and scalable blockchain-based trust management protocol with mobility support in massively distributed IoT systems. In our protocol, mobile smart objects disseminate trust information on service providers to the blockchain. Thus, all the objects will have a global view on each service provider in the architecture, which speeds up the trust evaluation process. In addition, our protocol is resilient against the most known malicious attacks such asbad-mouthing,ballot-stuffing, andcooperative attacks. We confirm the efficiency of our proposal through theoretical analysis and extensive simulations. Finally, we show that it outperforms existing solutions, especially in terms of scalability, mobility support, communication, and computation costs.

5G is the fifth generation technology standard for cellular networks, which is already being deployed and offering many advantages for communications, such as the high degree of flexibility, the possibility of management and control of systems and resources, thus allowing to host and execute services in one or more distinct network slices, etc. Non-orthogonal multiple access (NOMA) is being considered as a key enabling technique for 5G cellular systems. The earlier generations are becoming less practical with the emergent requirements such as very high spectral efficiency, very low latency, supporting diverse quality of services (QoS). Researchers are working on different techniques to fulfill these requirements, and NOMA is one of the essentials due to its capacity to increase spectrum efficiency. NOMA schemes exploit channel gain differences to serve multiple users concurrently, by superimposing multiple users' information signals at the transmitter side. In this paper, we review, analyze and classify recent papers on resource allocation problem in 5G networks, and then we discuss the proposed solutions. In particular, the papers are classified into two main categories: power/energy-efficient and rate-optimal. For each paper, the objective, optimization method and main results are discussed. Finally, some open challenges are highlighted.

Fog computing is a new paradigm which extends cloud-computing services into the edge of the network. However, this new architecture creates new challenges in relation with data security and privacy. In fact, due to the emergence of data externalization technologies, as Fog computing, privacy has become a major concern for all users. There have been several proposals in the literature that treated privacy issues using existing anonymization techniques, but few of them considered accountability service. Whereas, when security systems do not adopt accountability mechanisms, full anonymity may encourage users to act maliciously. In this paper, we propose a novel accountable privacy preserving solution for public information sharing in fog computing architecture. Based on signatures, our scheme allows fog servers to authenticate any user in the system without violating its privacy. In case of users' misbehavior, our architecture considers an authority that is able to trace any user in the system. Our solution ensures privacy preserving and accountability services in a completely distributed manner, without resorting permanently to the authority. Finally, we show through experimentation that our solution outperforms existing solutions.

Nowadays Cloud computing services are becoming a necessity due to the huge evolution of internet technology and the need of outsourcing data into external servers. However, data externalization introduces several issues. Access control is one of the most challenging issues, it must ensure data confidentiality through enforced and flexible access policies. Revocation is one of the important tasks of access control process, generally, it consists on banishing some roles from users on the system. Ciphertext-policy attribute-based encryption is a promising cryptographic method. It provides fine-grained access which makes it very useful in case of group-sharing applications. However, the challenge of this method is the revocation. Through few past years, there have been many proposed revocation solutions on ABE, but these solutions still lack of efficiency. In this paper, we propose an access control mechanism with an immediate and efficient attributes or users revocation. Our solution uses CP-ABE for a flexible and fine grained access while it provides collusion resistance, supports scalability and does not require neither rekeying mechanisms nor keys redistribution.