Protecting the cyber-physical systems (CPSs) against malicious attacks like false data injection (FDI) attacks is an enormous challenge. An FDI attacker may gather sufficient information about the system settings and design stealthy FDI attacks to degrade the CPS performance. Therefore, the early detection of FDI attacks is vital in CPSs. This paper addresses the detection of FDI attacks when the attacker simultaneously compromises the physical system, sensor measurements, and actuator inputs of the CPS. We express the attack detection as a partially observable Markov decision process (POMDP) problem and solve it using Chi-square (χ2) measurements with a sliding observation window and a reinforcement learning (RL) technique. We design an RL-based FDI attack detection mechanism and the reward parameters used in the training phase. The reward parameters are based on a state recognition rate. The effectiveness of the proposed detection mechanism is evaluated through numerical examples of different CPSs.

An attacker can modify the actuator inputs, sensor observations, and state of the physical system in a cyber-physical system (CPS), causing errors in the system’s proper functioning. It is crucial to investigate a CPS in the presence of all possible attack types to make it resilient. For this, all possible attack sequences must be known. This article focuses on designing false data injection (FDI) attacks on the physical system, actuator input, and sensor measurement in a CPS, individually and in their combined locations. Each attack sequence follows a Gaussian distribution. We study a discrete linear time-invariant (LTI) CPS with a single sensor and actuator. The system also includes a Kalman filter and a Chi-square ( \(\chi^{2}\) ) detector. With a \(\chi^{2}\) detector and possible known system parameters, we have proposed seven types of FDI attacks at vulnerable locations based on Kullback-Leibler (KL) divergence. The attacker can remain undetected by carefully planning the attack sequences. The attack increases the state estimation error (SER), and degrades the system’s proper operation. The effect of attacks on the detection result and the difference between SER with and without attack is simulated through two examples from the LTI system.

Internet of Things (IoT) and Blockchain integration are having a huge impact on the future of technological progress. IoT has progressed from an emerging notion to a widely used technology, shaping the future of digital connectivity. With billions of networked IoT devices generating large amounts of data, efficient data management becomes critical. Blockchain has been explored extensively to enhance security in IoT networks however, its scalability limitations become evident when handling large-scale deployments. Sharding is recognized as a promising approach to improve blockchain scalability by partitioning the network into multiple independent groups. These groups, called shards, process transactions in parallel, increasing throughput while reducing communication, computation, and storage over-head. Despite its advantages, many existing blockchain sharding models rely on static algorithm, which fail to accommodate the dynamic nature of blockchain networks. Factors such as variable node involvement and possible security concerns present problems that static sharding cannot solve. To address these restrictions, deep learning provides a strong solution for dynamic and multidimensional sharding in blockchain-based IoT systems. Deep learning, with its capacity to understand complex patterns and adapt to changing network circumstances, can improve the efficiency, security, and scalability of blockchain-powered IoT networks. This article proposes a deep reinforcement learning-based dynamic shards in blockchain IoT applications to overcome scalability difficulties.

ABSTRACT Timely delivery of critical messages is paramount for road safety in Vehicular Ad‐hoc NETworks (VANETs), making effective resource block allocation crucial. Proper allocation of radio resources ensures these critical messages are successfully delivered. VANETs can incorporates widespread cellular networks for high reliability and low latency communication. The challenge lies in prioritizing messages of varying importance generated randomly in the dynamic VANET environment. This paper introduces a methodology for Vehicle‐to‐Vehicle (V2V) communication, utilizing base stations to allocate resource blocks. The proposed research work considers ProSe Per‐Packet Priority (PPPP), distance between transmitter and receiver, and message size for prioritizing VUE pairs through the Analytic Hierarchy Process (AHP). A novel algorithm is introduced to maximize the information value of the overall network while optimizing resource allocation, thus, prioritizing the assignment of resource blocks to vehicle users with a higher Signal‐to‐Noise Ratio (SNR). Simulation results show that the information value of ranked VUE pairs is superior compared to reverse‐ranked VUE pairs and existing state‐of‐the‐art algorithms. Furthermore, compared to the existing method, the proposed approach significantly reduces the delay experienced by safety‐critical messages.

Cyber-physical systems (CPSs) are closed-loop feedback systems that efficiently control the physical processes through cyber-systems. Security threats like false data injection (FDI) attacks are increasing in CPSs. FDI attacks disrupt the system's performance by modifying the original data. It is necessary to have mechanisms in place to defend against FDI attacks and enhance security in CPSs. This paper proposes a two-stage defense scheme against an FDI attack that can simultaneously modify the data in physical and cyber systems. The first stage is attack detection using a watermarking technique with a Chi-square detector. The second stage is attack mitigation, in which we generate a time-varying compensation signal to reduce the attacker's effect and propose a control scheme. We demonstrate the simulation results of the proposed defense scheme against the FDI attack using an unmanned aerial system (UAS) as an example.

Abstract Containers in cloud computing provide a logical packaging technique for applications to be isolated from the computing environment in which they actually execute, allowing for efficient sharing of memory, processor, storage, and network resources at the Operating System (OS) level. Since they are so compact, container‐based clouds have recently gained significant popularity. In order to maximize resource usage and minimize energy consumption, the container consolidation technique is widely employed in the cloud environment. This work introduces container consolidation in cloud computing that exploits Fractional Pelican Hawks Optimization (FPHO). In Container as a Service (CaaS) model, containers are placed in the Virtual Machines (VMs), and virtual machines are hosted in Physical Machines (PMs) or servers. The proposed method for container consolidation consists of two modules, namely, the host status module and the consolidation module. In the host status module, the PM's load is predicted using Long Short Term Memory (LSTM) and checked whether the PM is overloaded or underloaded using a threshold. If it is overloaded, the container selection algorithm is performed, and the migration list is also generated. In the consolidation module, the created migration list which is employed for the destination list to be created by an overloaded destination selector. In the same way, the underloaded list is also generated by the underloaded destination selector. Finally, the container and VM migration is carried out by considering the multi‐objectives such as predicted load, migration cost, resource utilization, energy consumption, network, and bandwidth which are optimally selected by the proposed FHPO. Here, FHPO is the combination of Fractional Pelican Optimization (FPO) and Fire Hawk Optimizer (FHO). The designed model achieved the measures with minimum energy consumption, resource utilization, Service Level Agreement (SLA), and Makespan as 0.066, 0.019, 0.054, and 0.066, respectively for setup one.

Cloud computing is a revolutionary paradigm, which allows applications to run in a virtualized environment. The application runs on a virtual cloud resource makes the system scalable and cost-efficient. Noticeably many applications, such as healthcare systems, video streaming, Internet of Things (IoT) running in the cloud, are real-time in nature, i.e., these applications demand responses within a particular time limit, i.e., deadline. To meet the requirement of such applications, a Cloud Service Provider (CSP) must have a sufficient number of cloud resources (virtual machines). Further, the ever-growing demand for applications forces a CSP to deploy more and more cloud resources. Inevitably, the massive count of cloud resources in a cloud data center consumes a tremendous amount of energy. Specifically, it becomes cumbersome to offer services to deadline-sensitive tasks while minimizing energy consumption. An efficient task scheduling is an attractive way to reduce energy usage while ensuring satisfactory services for cloud users. Learning Automata (LA) is a reinforcement-based adaptive decision-making unit that learns and selects the best action from a set of actions applied in a dynamic environment. Similar to LA, in task scheduling, the best task and virtual machine combinations are chosen from a set of available combinations. In this context, this paper implemented the LA technique to solve a bi-objective deadline-sensitive task scheduling problem which includes minimization of energy consumption and makespan. At first, a learning automata-based scheduling framework is designed for deadline-sensitive tasks in the cloud. Later, a scheduling algorithm, namely, the LA-based Scheduling (LAS) algorithm, is proposed. The LAS algorithm exploits the heterogeneity of tasks and virtual machines (VMs) while guaranteeing the task’s deadline. Extensive simulation is carried out to designate the effectiveness and applicability of LAS for deadline-sensitive task scheduling in the heterogeneous cloud environment.

The evolution of cloud computing facilitates applications with varying demands to operate in a virtualized environment. For instance, applications like the healthcare system, video streaming, Internet of Things (IoT) that are moving to the cloud, demand responses within a particular time limit, i.e., deadline. However, the cloud computing system consumes a considerable amount of electric energy while providing services to these type of applications, which in turn contribute to the high operational cost. Specifically, it becomes cumbersome to offer services to deadline sensitive task while minimizing energy consumption. In this regard, efficient task scheduling is an attractive way to cut down energy usage while ensuring satisfactory services for cloud users. In this paper, the task scheduling problem is considered as a bi-objective minimization problem which includes minimization of energy consumption and makespan. First, we proposed a novel learning automata-based scheduling framework for deadline sensitive tasks in the cloud. Learning automata (LA) is an adaptive decision-making unit that helps the scheduler to select the best responses. Later, the LA-based Scheduling (LAS) algorithm is introduced which exploits the heterogeneity of tasks and virtual machines (VMs) while ensuring the timing requirements of the tasks. Extensive simulation is carried out to designate the effectiveness and applicability of LAS for deadline sensitive task scheduling in the heterogeneous cloud environment.

Neural networks are extensively used for determining different characteristics of optical burst switching networks. The main disadvantage of optical burst switching network is burst drop and burst contention, which occurs because of burst getting blocked. Using neural network approaches, blocking probability can be pre-determined for the upcoming traffic. In this paper, Log-incremental modified extreme learning machine trained generalized radial basis function neural network (MELM-GRBFNN) model is used for training and predicting burst contention or burst blocking probability. From the obtained results, it is inferred that the prediction accuracy of our proposed model is more accurate and faster than the contemporary approaches. It is observed that our proposed method is competent in predicting the burst blocking probability with higher accuracy and indicates a reduction in the burst loss. Thus, it will help network designers to have a preliminary idea about the performance of the network model under specific configurations.

Node replication attack possesses a high level of threat in wireless sensor networks (WSNs) and it is severe when the sensors are mobile. A limited number of replica detection schemes in mobile WSNs (MWSNs) have been reported till date, where most of them are centralized in nature. The centralized detection schemes use time-location claims and the base station (BS) is solely responsible for detecting replica. Therefore, these schemes are prone to single point of failure. There is also additional communication overhead associated with sending time-location claims to the BS. A distributed detection mechanism is always a preferred solution to the above kind of problems due to significantly lower communication overhead than their counterparts. In this paper, we propose a distributed replica detection scheme for MWSNs. In this scheme, the deviation in the distance traveled by a node and its replica is recorded by the observer nodes. Every node is an observer node for some nodes in the network. Observers are responsible for maintaining a sliding window of recent time-distance broadcast of the nodes. A replica is detected by an observer based on the degree of violation computed from the deviations recorded using the time-distance sliding window. The analysis and simulation results show that the proposed scheme is able to achieve higher detection probability compared to distributed replica detection schemes such as Efficient Distributed Detection (EDD) and Multi-Time-Location Storage and Diffusion (MTLSD).

Wireless sensor networks are susceptible to various kinds of security threat. Node replication attack is one of them. In this attack, an adversary captures a node and deploys a number of replicas in the network to launch further attacks. Detection of replica is a challenging task especially when the nodes are mobile. Only a few techniques are reported in the literature for detecting replicas in mobile wireless sensor networks. They are mostly centralized detection mechanisms, where the base station plays a major role in clone detection. In this paper, we propose a distributed replica detection mechanism called energy-based replica detection. Every node in the network acts as a monitoring node to a set of nodes in the network. Replicas are detected on the basis of the residual energy of nodes. A conflict in the timestamp-residual energy pair of a node is detected as clone by its monitoring node. We have simulated and compared the proposed scheme with Efficient Distributed Detection and Multi-Time-Location Storage and Diffusion scheme. We observed that the proposed scheme has higher detection probability, and lower communication and storage overhead. Copyright © 2014 John Wiley & Sons, Ltd.