Industrial Internet of Things (IIoT) opens up a challenging research area towards improving secure data sharing which currently has several limitations. Primarily, the lack of inbuilt guarantees of honest behavior of participating, such as end-users or cloud behaving maliciously may result in disputes. Given such challenges, we propose a fair, accountable, and secure data sharing scheme, FairShare for IIoT. In this scheme, data collected from IoT devices are processed and stored in cloud servers with intermediate fog nodes facilitating computation. Authorized clients can access this data against some fee to make strategic decisions for improving the operational services of the IIoT system. By enabling blockchain, FairShare prevents fraudulent activities and thereby achieves fairness such that each party gets their rightful outcome in terms of data or penalty/rewards while simultaneously ensuring accountability of the services provided by the parties. Additionally, smart contracts are designed to act as a mediator during any dispute by enforcing payment settlement. Further, security and privacy of data are ensured by suitably applying cryptographic techniques like proxy re-encryption. We prove FairShare to be secure as long as at least one of the parties is honest. We validate FairShare with a theoretical overhead analysis. We also build a prototype in Ethereum to estimate performance and justify comparable results with a state-of-the-art scheme both via simulation and a realistic testbed setup. We observe an additional communication overhead of 256 bytes and a cost of deployment of 1.01 USD in Ethereum which are constant irrespective of file size.

Recently collaborative learning is widely applied to model sensitive data generated in Industrial loT (1IoT). It enables a large number of devices to collectively train a global model by collaborating with a server while keeping the datasets on their respective premises. However, existing approaches are limited by high overheads and may also suffer from falsified aggregated results returned by a malicious server. Hence, we propose a Scal-able, Privacy-preserving and veRIfiable collaboraTive lEarning (SPRITE) algorithm to train linear and logistic regression models for IloT. We aim to reduce burden from resource-constrained IloT devices and trust dependence on cloud by introducing fog as a middleware. SPRITE employs threshold secret sharing to guarantee privacy-preservation and robustness to IloT device dropout whereas verifiable additive homomorphic secret sharing to ensure verifiability during model aggregation. We prove the security of SPRITE in an honest-but-curious setting where the cloud is untrustworthy. We validate SPRITE to be scalable and lightweight through theoretical overhead analysis and extensive testbed experimentation on an IloT use-case with two real-world industrial datasets. For a large-scale industrial setup, SPRITE records 65% and 55% improved performance over its competitor for linear and logistic regressions respectively while reducing communication overhead for an IloT device by 90%.

The advent of Industrial Internet of Things (IIoT) along with cloud computing has brought a huge paradigm shift in manufacturing industries resulting in yet another industrial revolution, Industry 4.0. Huge amounts of delay-sensitive data of diverse nature are being generated, which need to be locally processed and secured because of their sensitivity. However, the low-end Internet of Things devices are unable to handle huge computational overheads. In addition, the semi-trusted nature of cloud introduces several security concerns. To address these issues, this article proposes a secure fog-based IIoT architecture by suitably plugging a number of security features into it and by offloading some of the tasks judiciously to fog nodes. These features secure the system alongside reducing the trust and burden on the cloud and resource-constrained devices, respectively. We validate our proposed architecture through both theoretical overhead analysis and practical experimentation, including simulation study and testbed implementation.

Two of the most dominant challenges in post disaster emergency resource allocation are: 1) understanding the exact utility, i.e., exigency, of emergency resources and 2) collecting and transmitting the need for these resources to the control station from where resources are allocated. Measuring the utility of resources with precision becomes tricky in a dynamic post disaster scenario, where demands are constantly evolving and supplies trickle in at an uncertain rate. Moreover, collection and transmission of resource needs of far-flung areas are easier said than done owing to the post disaster disruption of communication infrastructure. These result in the ad-hoc allocation of emergency resources to the shelters. In this paper, we first derive a utility function for dynamically enumerating the shelter specific utility of each emergency resource. Subsequently, we propose an opportunistic knowledge sharing scheme for gathering and disseminating resource needs to the control station using a smartphone-based delay tolerant network. Finally, based on these opportunistically transmitted needs, we formulate a utility driven optimal resource allocation model which minimizes overall resource deficit and total resource deployment time. The proposed system optimally assigns constrained emergency resources to different shelters, so that high-utility resources are deployed fast. The effectiveness of the proposed system is evaluated using ONE simulator and LINGO optimization modeling tool. Exhaustive simulation is done to evaluate the comparative performance of our system with a number of competing schemes. Results show that our system outperforms all these schemes in a fully connected scenario. It is also observed that even in an intermittently connected environment, the performance of our system is almost at par with the competing schemes.

Message authentication codes (MACs) are classically used for preventing unauthorized and corrupted messages from being forwarded in a network. However, inherent energy limitations of wireless sensor networks (WSNs) make the application of most of the state-of-the art MACs unaffordable due to their large computation overhead. Therefore in this paper, in order to cope with this challenging concern, we have proposed a lightweight hash based symmetric key message authentication code. The primary focus is on making the algorithm lightweight so that on using it in achieving secured communication in energy starved networks like WSNs, the resource constrained nodes can successfully run the algorithm. Detailed security analysis shows that LMAC also thwarts passive attack as well as active attack. Finally the comparative usability of the MAC in the said application domain is worked out and that shows the dominance of LMAC over several state-of-the-art MACs. We claim that on an average LMAC requires 61% less overhead compared to its competitors.

In Delay-Tolerant Networks (DTNs), nodes forward data opportunistically upon contact because of irregular connectivity and low network-denseness. In this type of networks, energy plays an important role in successful data delivery and enhancement of network lifetime. To address the problem of selecting appropriate relay nodes for data forwarding, recently researchers are using concepts of social networks in the design of DTN routing schemes. This is because of its ability to make forwarding decisions based on the knowledge of long-term and more stable social characteristics of the nodes. In this paper, we propose an energy-aware relay selection scheme to support multiple multicast sessions based on an existing social-based single-data multicast scheme. The proposed scheme selects relays for delivering data to the destinations based on the centrality, one of the important social network characteristics, as well as on the residual energy of the DTN nodes. Our proposed scheme is shown to have polynomial-time complexity. We claim that our scheme supports higher number of multicast sessions thereby maintaining longer lifetime of the network while achieving required data delivery ratio compared to a prominent single-data multicast scheme. Our claim is validated through quantitative performance analysis.

In wireless sensor networks the issue of preserving energy requires utmost attention. One primary way of conserving energy is judicious deployment of sensor nodes within the network area so that the energy flow remains balanced throughout the network and prevents the problem of occurrence of energy holes. Firstly, we have analyzed network lifetime, found node density as the parameter which has significant influence on network lifetime and derived the desired parameter values for balanced energy consumption. Then to meet the requirement of energy balancing, we have proposed a probability density function (PDF), derived the PDF's intrinsic characteristics and shown its suitability to model the network architecture considered for the work. A node deployment algorithm is also developed based on this PDF. Performance of the deployment scheme is evaluated in terms of coverage-connectivity, energy balance and network lifetime. In qualitative analysis, we have shown the extent to which our proposed PDF has been able to provide desired node density derived from the analysis on network lifetime. Finally, the scheme is compared with three existing deployment schemes based on various distributions. Simulation results confirm our scheme's supremacy over all the existing schemes in terms of all the three performance metrics.

ABSTRACT Wireless sensor networks (WSNs) are susceptible to various attacks, and these attacks pose a major threat to the normal functioning of WSNs. Jamming is one such attack, which affects network operations by blocking frequencies. Hence, the need of defending such attack is of utmost importance. This paper proposes three variants of a data forwarding scheme using multi‐level multi‐tier architecture, which employs judicious surfing of a pair of channels to defend jamming attack in WSNs. Variant I defends the attack at the cost of an affordable delay. Improvement of the scheme in minimizing data transmission time is made in variant II. Further improvement in transmission time is made in variant III and is made more realistic by introducing sensing capability at all tiers, thereby increasing network coverage. Simulation is performed to establish each of the successor's variant's improvement over its predecessor. The performance of the scheme in terms of area coverage, packet delivery ratio, and control message overhead is compared with an existing jamming‐defending scheme for infrastructure‐based wireless network. Finally, comparison results establish that although the proposed scheme is equivalent to its infrastructure‐based counterpart in terms of coverage, it provides a more lightweight solution and maintains improved performance in terms of packet delivery ratio. Copyright © 2013 John Wiley & Sons, Ltd.

Continuous monitoring of the environment in coal mines is an important application area in wireless sensor networks. It is very essential to know the conditions that exist in coal mines for ensuring safe working inside the mines. Based on the characteristics of room-and-pillar mining technique and the requirements for mine safety monitoring, we propose a heterogeneous node deployment strategy. To obtain real time and comprehensive monitoring, nodes are deployed in some strategic locations of the underground coal mine. We claim that the strategic locations are chosen in such a way that the proposed deployment scheme ensures 1-coverage and 3-connectivity. The claim has been justified by supporting lemma and proof. The present work achieves enhanced network lifetime by ensuring efficient energy usage amongst the nodes thereby prolonging network lifetime. Further, exhaustive simulation is performed considering a realistic routing protocol to substantiate our claim of energy efficiency and subsequent enhancement of network lifetime. Finally the scheme is compared with two existing deployment schemes. The results confirm our scheme's supremacy over both the existing schemes in terms of the performance metrics such as energy efficiency and network lifetime.

This work presents an efficient scheme for processing queries on location dependent data in a cellular mobile environment. It is developed around the theory of cellular automata (CA) which act as a search engine and implements computation for query processing that either can be done at the network base station or at the mobile unit depending on the requirement specified by the service provider. The scheme also ensures efficient handling of disconnection problem due to hand-off. The experimental results confirm that the overhead & delay, query processing, in the proposed scheme is much less than that of the existing schemes.

Benes network, being a back-to-back connection of two Baseline networks, the method for fault diagnosis for the class of nonredundant networks, as elucidated in our previous work, can be directly mapped on the two nonredundant networks. The individual results from these two networks can be combined to construct a comprehensive algorithm for the Benes network to diagnose single fault.