Industrial Systems and Industrial Data Privacy -- A Review

Industrial Systems and Industrial Data Privacy -- A Review | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 30 January 2025 V1 Latest version Share on Industrial Systems and Industrial Data Privacy -- A Review Author : Bahaa Eltahawy 0000-0001-6372-7547 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.173822597.78031676/v1 421 views 150 downloads Contents Abstract Keywords Cloud Computing Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract The rise of industrial systems driven by recent advances in engineering and data science has significantly changed the landscape of manufacturing and production. Equipped with modern tools and capabilities, industrial systems can streamline processes, enhance production, analyze complex scenarios, and support decision-making. Central to these systems is industrial data, which provides the insights and means necessary to drive operations and achieve production objectives. Given its critical value, protecting industrial data from potential risks is essential for ensuring consistency, utility, and competitiveness. While various studies have focused on security factors, the literature addressing industrial data privacy remains limited. Recognizing this gap and the importance of both industrial systems and data privacy, this study thoroughly explores these topics. First, industrial systems are examined, highlighting their prevalent types and establishing a foundation for understanding their distinctive features. Next, 34 selected studies on industrial data privacy are reviewed, discussing its significance, current challenges, and potential solutions. The study identifies 10 common types of industrial systems and their shared characteristics. Additionally, it presents 15 definitions and contexts, proposing an inclusive definition that aligns with modern industrial systems. The study also defines industrial data and identifies eight contexts associated with industrial data privacy, providing a comprehensive review of each. Finally, it highlights and recommends a range of solutions including operational and technical means for protecting industrial data. Overall, the findings underscore the pressing need to prioritize industrial data privacy and address it more closely in both research and practice. Industrial Systems and Industrial Data Privacy – A Review Bahaa Eltahawy* Computer Science, School of Technology and Innovations, University of Vaasa Wolffintie 32, 65200 Vaasa, Finland [email protected] *Corresponding author Abstract. The rise of industrial systems driven by recent advances in engineering and data science has significantly changed the landscape of manufacturing and production. Equipped with modern tools and capabilities, industrial systems can streamline processes, enhance production, analyze complex scenarios, and support decision-making. Central to these systems is industrial data, which provides the insights and means necessary to drive operations and achieve production objectives. Given its critical value, protecting industrial data from potential risks is essential for ensuring consistency, utility, and competitiveness. While various studies have focused on security factors, the literature addressing industrial data privacy remains limited. Recognizing this gap and the importance of both industrial systems and data privacy, this study thoroughly explores these topics. First, industrial systems are examined, highlighting their prevalent types and establishing a foundation for understanding their distinctive features. Next, 34 selected studies on industrial data privacy are reviewed, discussing its significance, current challenges, and potential solutions. The study identifies 10 common types of industrial systems and their shared characteristics. Additionally, it presents 15 definitions and contexts, proposing an inclusive definition that aligns with modern industrial systems. The study also defines industrial data and identifies eight contexts associated with industrial data privacy, providing a comprehensive review of each. Finally, it highlights and recommends a range of solutions including operational and technical means for protecting industrial data. Overall, the findings underscore the pressing need to prioritize industrial data privacy and address it more closely in both research and practice. Keywords: Industrial Systems, Industrial Data, Privacy, Data Protection, Industrialization, Industry 4.0, Emerging technologies Introduction Overview In the past two decades, manufacturing and production have undergone a remarkable transformation, driven by the convergence of advanced technologies and shifting demands that have reshaped industries globally (Brennan et al. , 2015). Thanks to rapid technological innovations and the strategic utilization of industrial data, the dynamic interaction between systems, customization, and process optimization have revolutionized the landscape of manufacturing and production. Specifically, advances in engineering, Information Technology (IT), and Operational Technology (OT) have converged to drive a new shift in industrial systems, encompassing comprehensive frameworks, specialized tools, and capabilities designed to meet the challenges of modern manufacturing (Agrawal and Berm, 2015; ElMaraghy et al., 2021). With the introduction of Industry 4.0 (I4.0) in 2011 and the integration of the Internet of Things (IoT), data analytics, and automation technologies, industrial systems became more interconnected and intelligent (Yang, 2017; Zhong et al., 2017). This resulted in higher precision, customization, adaptability, speed, and efficiency, as seen in smart factories where manufacturing is fully digitalized and systems continuously collect, process, and share data for real-time monitoring, predictive maintenance, and adaptive manufacturing (ElMaraghy et al., 2021). The result is a very dynamic and responsive industrial ecosystem capable of meeting complex demands and adapting to abrupt changes while minimizing interruptions, controlling costs, and ensuring product quality. Industrial data plays a fundamental role in this context, serving as the key enabler for the intelligence in industrial systems (Bordeleau et al., 2018). Only through the utilization of industrial data generated by sensors, machinery, interconnected devices, as well as market and supply chain signals and processes, has value been created and manufacturing transformed (Wang, 2018). Within the I4.0 paradigm, data is no longer merely a byproduct of industrial systems, but is rather a strategic asset (Martínez et al., 2021; Kayabay et al., 2022). It carries real-time metrics, measures, indicators, parameters, and control commands that are analyzed and used to provide operational and production insights (ElMaraghy et al., 2021). As the role of industrial data has grown in significance, associated risks have also increased, including potential cyberattacks, data breaches, and the exploitation of sensitive information (Raptis et al., 2019; Corallo et al., 2021; Haleem et al., 2022). These can lead to several potential consequences, such as operational malfunctions, production disruptions, compromised systems, intellectual property theft, and even impacts on personal data rights (Onik et al., 2019). Several trials have been conducted to address these issues, e.g., (Waidner and Kasper, 2016; Timan and Mann, 2021; Sun et al., 2021; Guan, 2023), as well as the studies reviewed in Section 4. However, these efforts either addressed industrial data topics separately or focused on aspects and measures of cybersecurity and governance frameworks without giving sufficient attention to the actual value of data and its privacy while being processed and shared. Accordingly, more comprehensive approaches that integrate privacy practices into industrial data are needed, as protecting and maintaining the integrity of industrial data is essential for ensuring consistency, utility, and competitiveness. Motivation, Research Questions, Contributions, and Limitations Motivated by the rise of industrial systems and the significance of industrial data, this study investigates the existing literature to answer the central research question: “What is the current status of industrial data privacy?” The question is further refined using the Mutually Exclusive and Collectively Exhaustive (MECE) framework (Chevallier, 2016) into the following three sub-questions: 1) “What are industrial systems?” 2) “What is industrial data?” and 3) “What challenges and solutions exist regarding industrial data privacy?” The main contributions of this study are threefold: 1) Proposing a definition of industrial systems that aligns with their current and modern capabilities, features, and characteristics; 2) The distinction of being one of the first – if not the only – dedicated endeavors to comprehensively cover the topic of industrial data privacy from all relevant perspectives in the existing literature; and 3) Providing eight themes related to industrial data privacy along with a set of recommended solutions. Finally, this study is based on desk research, thus only reviewing available literature without producing any primary data or solutions of its own. Research Organization The remainder of this work is structured as follows: Section 2 defines industrial systems, outlines their common types and characteristics, and reviews industrial data attributes and classifications. Section 3 describes the methodology employed in this study. Section 4 provides a comprehensive literature survey on industrial data privacy, highlighting identified risks and solutions. In Section 5, findings are synthesized and discussed. Finally, Section 6 concludes and provides directions for future work. Background – Industrial Systems Industrial Systems – Definitions First, finding a unanimous and precise definition of the term “industrial systems” was challenging, primarily due to the term’s broad applicability and contextual nature across various sectors. In Table 1, the most common definitions of the term “industrial systems” are collected, highlighting their contextual backgrounds. Table 1. Industrial systems definitions and their contexts Contexts Definition Manufacturing; Services; Value-creation “An industrial system includes the context, resources, activities, processes, actors, and interdependencies that support the creation and delivery of products and services. A clearer understanding of industrial systems – a holistic view – can identify those ‘levers’ which are available to generate and, crucially, capture value” (The Royal Academy of Engineering, 2012) Energy; Operations An “industrial system means the whole or any part of an electric system primarily intended to serve one or more industrial operations of which the system forms a part” (Jamil and Pearce, 2022) Internet of Things; Networks; Machinery “The industrial systems of the future are complex systems composed of vast numbers of devices interacting with each other and with enterprise systems.” (Holler, 2014) Management “An industrial system is a collection of interrelated elements brought together to achieve a specific objective of meeting product or service goals” (Badiru et al. , 2007) Data mining; Manufacturing “The industrial system is a typical real-time control and real-time information processing system” (Luo, 2022) Industry 4.0 “An industrial system is an ensemble of parts connected in a networked fashion, which show a peculiar behavior that is not observable when the parts are considered separately” (Urbani, 2021) Business “The modern industrial system is a concatenation of processes which has much of the character of a single, comprehensive, balanced mechanical process” (Veblen, 2003) Environment management “An industrial system is an ecological system” (Levine, 2003) Industrial symbiosis “The regional industrial system (RIS)” is “a more or less stable collection of firms located in proximity to one another, where firms in principle can develop social and material/energy connections as a result of that proximity” (Boons et al. , 2011) Marketing; Production “The industrial system is a network of firms engaged in the production, distribution and use of goods and services through which lasting business relationships are established, developed and maintained” (Whitelock, 2002) Behavioral management; Accounting “The industrial system … is a very complex multi-loop and interconnected system … Decisions are made at multiple points throughout the system. Each resulting action generates information that can be used at several but not all decision points. This structure of cascaded and interconnected information-feedback loops, when taken together, describes the industrial system.” (Riahi-Belkaoui, 2001) Informatics; Data “A large-scale industrial system is a networked information system, where the raw data sampled at the lowest device level flows up to upper-layers. Various data acquisition and processing tasks are carried out at different layers for different purposes.” (Dai and Gao, 2013) Private data; Internet of Things; Blockchain “An industrial system is a loosely distributed organization” (Saiyu et al. , 2020) Simulation; Modeling “Eco-industrial system is a typical complex adaptive system that generates intricate patterns with given constraints … An eco-industrial system always tries to find an optimal process to obtain maximized flux under given constraints.” (Huo and Chai, 2008) Manufacturing; Business “An industrial system is a highly complex system … It consists of many interacting and interconnected autonomous entities that are continuously adapting, while new entities are added and old entities are removed. As a result of this complexity, it is difficult – if not impossible – to predict the development of the industrial system.” (Bas, 2017) As seen in Table 1, the term “industrial systems” varies significantly depending on the sectors considered and their specific characteristics. Additionally, common themes can be identified across the given definitions, including complexity, adaptability, purpose, information-driven operations, networked and interlinked architecture, as well as an ecological perspective. Industrial Systems – Types and Characteristics Second, industrial systems are not limited to a single form or structure; rather, they come in various types and sizes depending on the sector in which they are deployed and the purpose they are intended to fulfill, as highlighted in the definition proposed earlier. Below is a list of the most common industrial systems identified while conducting this research. It is important to note that this list is not exhaustive, as industrial systems continuously evolve and extend beyond these examples. 1. Industrial control systems (ICS) are integrated infrastructures to control industrial systems distributed over large geographical areas and locations. These include networks, sensor devices, and controllers to automate and operate industrial tasks and processes effectively. ICS are either Supervisory Control & Data Acquisition (SCADA), Distributed Control Systems (DCS), or hybrid systems that combine the best features of both systems (Macaulay and Singer, 2011). 2. Supervisory control and data acquisition (SCADA) refers to the centralized systems that control production infrastructures. SCADA is frequently used interchangeably with process control and ICS; however, the distinction may lie in the observation that SCADA systems are considered to support the coordination of infrastructures rather than controlling the discrete elements of these infrastructures. ICS encompasses both coordination and control functions. 3. Distributed control systems (DCS) refer to systems in which the controller elements are distributed rather than centralized – as in SCADA – with each component and discrete subsystem managed by one or more controllers. 4. Programmable Logic Controller (PLC) is the control component of the ICS that provides process management. PLC provides supervisory, remote access, and control to devices such as actuators and sensors (Bhardwaj et al. , 220). 5. Human Machine Interface (HMI) provides a graphical user interface (GUI) application that facilitates the interaction between hardware, control system, and operators (Bhardwaj et al. , 220). 6. Safety Instrumented Systems (SIS) are hardened ICS elements designed for high reliability and safety in the event of system failure. SIS includes functional elements that contribute to operational safety and risk management, often sharing technical architectures and features with more general-purpose ICS (Macaulay and Singer, 2011). 7. Industrial Automation and Robotics is the control of machinery and processes used in various industries by autonomous systems through technologies like robotics and computer networks (ElMaraghy et al., 2021; Campilho and Francisco, 2023). 8. Machine Vision (MV) refers to the technology and methodologies used for imaging-based autonomous inspection and analysis in various applications. MV is used for material inspection, object recognition, pattern recognition, electronic component analysis, signature recognition, optical character recognition, and money recognition (Javaid et al. , 2022). 9. Additive Manufacturing (AM) , a subset of Adaptive Manufacturing, is the process of joining materials to create parts from 3D model data, usually built in layers (Campbell et al. , 2011). 10. Industrial communication networks are the infrastructure that connects bottom field devices, such as sensors and actuators, with control devices like PLC, DCS, and SCADA to achieve industrial automation control, system interconnection and interoperability. These networks include fieldbus, Industrial Ethernet, industrial wireless networks, and other heterogenous networks (Liu et al. , 2022). 11. Industrial Edge Computing (IEC) is a system of micro data centers located at the edge of the network, close to or within factory premises. In edge computing, computing tasks are executed near the end users or devices in terms of geographical and network proximity. This enables levels of latency and throughput that are unattainable with cloud computing (Harmatos and Maliosz, 2021). 12. Industrial analytics (Industry 4.0 analytics , or Industrial Intelligence ) is an interdisciplinary field that bridges data science and industrial engineering, and it lies at the core of Industry 4.0 and the Industrial Internet of Things (IIoT). Industrial analytics aims at exploiting the business value of data by developing data-driven products, services, and processes (Gröger, 2022). 13. Artificial Intelligence (AI) is “the science and the engineering of making intelligent software systems and machines” (Wang, 2019). It is the “simulation system of collecting knowledge and information and processing intelligence … and disseminating it to the eligible in the form of actionable intelligence” (Grewal, 2014). 14. Industrial Internet of Things (IIoT) is “the extension and use of the Internet of Things (IoT) in industrial sectors and applications. A powerful emphasis on big data, machine-to-machine communication, and machine learning, the IIoT enables industries and enterprises to have better efficiency and credibility in their operations” (Perwej et al. , 2019). IIoT “is built for bigger ‘things’ than smartphones and wireless devices. It aims at connecting industrial assets, like engines, power grids and sensor to cloud over a network” (Helmiö, 2017). 15. Cyber-Physical Systems (CPS) are the integration of computing and physical processes. CPS are primarily characterized by the flow of information involving multiple heterogeneous physical systems (Liu et al. , 2017). 16. Manufacturing execution systems (MES) are IT tools commonly deployed in organizations involved in traditional manufacturing. An MES enables information exchange between the organizational level, commonly supported by an Enterprise Resource Planning (ERP), and the control systems for the shop-floor, usually consisting of several, different, highly customized software applications (D’Antonio et al. , 2017). 17. Energy Management Systems (EMS) are complex hardware and software systems that assist in monitoring and controlling power systems to minimize operating expenses, ensure a continuous power supply, and maintain an adequate operating margin to accommodate potential power outages. EMSs also perform essential functions such as data collection and maintenance for customer billing (Avramovic and Fink, 1995). 18. Quality Management Systems (QMS) are management systems comprising interconnected and interacting processes that work together to control and achieve an organization’s quality objectives (International Organization for Standardization, 2015; Paraschivescu, 2016). 19. Industrial Cybersecurity is “a set of practices, processes and technologies designed to manage cyber risks arising from the operation, processing, storage and transmission of information used in industrial organizations and infrastructures, using a people, process and technology perspective” (Werbińska-Wojciechowska and Winiarska, 2023). Industrial Cybersecurity is a “dedicated portfolio of technologies and services designed to protect operational technology layers and elements of industrial enterprises – including SCADA servers, HMIs, engineering workstations, PLCs, network connections and even engineers – without impacting on operational continuity and consistency of industrial processes” (Schwab and Poujol, 2018). As seen from the given examples, industrial systems share key characteristics and features that strongly distinguish them from consumer and commercial systems. Below is a brief list of the identified characteristics: Integrated infrastructure and connectivity: as of being composed of multiple interconnected physical components and software applications that work together within processes to achieve specific objectives. 1. Complexity: as of being interconnected with multiple components that have a wide range of functionalities and intricate processes. 2. Automation: as of utilizing software tools, programmable logic, and similar technologies and methods to efficiently and accurately execute tasks and streamline operations while adapting to changing conditions. 3. Distributed control and coordination: as of being distributed across large areas yet still functioning as a unified system. 4. Data-driven: as of being equipped with sensors and data acquisition systems to optimize processes and enhance decision-making 5. Advanced technologies and specialized networks: as of utilizing cutting-edge technologies with enhanced capabilities, throughput, and latency to enable efficient and reliable operations. 6. Focus on efficiency, reliability, and scalability: as of emphasizing the achievement of higher efficiency and reliability levels while enhancing seamless integration and robust scalability to ensure longevity and meet future demands. 7. Risk management and resilience: as of being designed to withstand unexpected events and disruptions, mitigate risks, and ensure continuous operations. By integrating the key characteristics and features identified above with the definitions and themes identified from Table 1, a comprehensive definition of “modern” industrial systems is suggested as follows: Industrial systems are sophisticated, interconnected networks of components, subsystems, and processes that utilize information, control systems, automation, and adaptive mechanisms to achieve specific objectives related to the production and delivery of goods and services, while emphasizing efficiency, reliability, scalability, and resilience through data-driven approaches. Industrial Data Finally, “w hat is industrial data?” is a fundamental question of this research, as the answer is essential for understanding the nature of industrial data and how to preserve its privacy. Fortunately, Xu et al. (2021) have clearly addressed this issue, providing a comprehensive classification of industrial data and its sources, as shown in Table 2. Table 2. Industrial data classifications and their sources (Adopted from Xu et al. (2021)) Data field Classification Sources R&D R&D design and Development data Computer-Aided Design (CAD); Simulation and analysis Computer-Aided Engineering (CAE); Industrial software development system; Industrial system testing tools; etc. Production Control information; Industrial control; Process parameters; and System log MES; PLC; SCADA; DCS; QMS; Working condition database; etc. Operation and maintenance Logistics and After-sales maintenance data EMS; Product logistics system; Product after-sales status tracking system; After-sales service management system; etc. Management System equipment asset information; Customer and product information; Product supply chain data, Business statistics, and Safety Systems Product Lifecycle Management (PLM); Supply Chain Management (SCM); QMS; Enterprise Resource Planning (ERP); Customer Relationship Management (CRM); Warehouse Management System (WMS); SIS; etc. External Data shared with other subjects Access to supply chain and collaborative R&D; etc. Platform Operation Data collection; Customer application data; Knowledge base repository; Analysis data; Configuration data; Application data; Technical and management data Production data; Monitoring data; Sensors; Customer platform-independent data; Knowledge base and cloud data; Reports, results and data analysis techniques; Configurations, user accounts, application services, and equipment; Industrial applications; Source codes, tools, test cases; etc. Enterprise management Customer and solution data; Business cooperation; Personnel; and Financial data Customer information; Behavior characteristics; Usage records; Customer service and maintenance records; Customization and deliverables; Agreements; Contracts; Employee information; Assets; Audit information; Financial statements; etc. Having defined industrial systems and industrial data, we proceed to addressing the issue of industrial data privacy in the following sections. Methodology In this work, we followed the systematic literature review practices of Okoli and Schabram (2015) along with the recommendations of Rowe (2014) and Schryen (2015) to identify and synthesize literature on industrial data privacy. First, three databases were considered: Google Scholar for its broad coverage, followed by Scopus and IEEEXplore for their specialized, focused content. The search strings used are: “Industrial data privacy”, “Privacy of industrial data”, and “Privacy of the industrial data”. Second, due to the limited number of results – less than 50 – all results were considered, regardless of the publication year. Table 3 shows search strings and their corresponding results. Table 3. Search strings and the corresponding results. # Search String Database Returned Results and description 1 Industrial data privacy Google Scholar Scopus IEEEXplore 22 – inclusive 3 – duplicated in Google Scholar 3 – duplicated in Google Scholar 2 Privacy of industrial data Google Scholar Scopus IEEEXplore 20 – inclusive 2 – duplicated in Google Scholar 2 – duplicated in Google Scholar 2 Privacy of the industrial data Google Scholar Scopus IEEEXplore 7 0 0 Finally, after removing duplicates, assessing quality and relevance, and considering peer-reviewed scientific content from conferences or journals (with three articles exempted from the peer-reviewed criterion due to their relevance) exclusively in English while excluding others from the scanning phase, 34 articles were considered for the final review stage. Figure 1 illustrates the adopted methodology. Image (image1.png) is missing or otherwise invalid. Fig. 1. The adopted research methodology Industrial Data Privacy – A Literature Survey Preliminary Insights 6. Timeline Figure 2 shows the frequency of reviewed literature covering the topic of industrial data privacy. As shown in the figure, no explicit research addressing this topic with the given search strings was found before 2017. However, since 2017, the topic has received increased attention, reflecting a growing awareness of its potential and significance. This can be explained in light of the increasing capabilities of industrial systems, their potential impacts, increased security and privacy concerns, and the growing recognition of the importance of industrial data. Image (image2.png) is missing or otherwise invalid. Fig. 2. Frequency of research covering the topic of industrial data privacy Keywords Figure 3 shows the frequency of keywords extracted from the reviewed literature. This serves as a preliminary tool to provide insight into the covered topics and their relevance. It is worth mentioning that the column Security Techniques differs from the more general column Security, as it includes keywords related to techniques, such as encryption, decryption, authentication, confidentiality, key generation, etc. Image (image3.png) is missing or otherwise invalid. Fig. 3. Frequency of keywords extracted Before proceeding with the literature survey and synthesis, the reviewed studies were categorized into relevant topics as shown in Table 4. In Table 5, the reviewed studies are mapped to the defined categories. Table 4. Reviewed studies’ categories and topics Category Topics C1: Privacy Data privacy, differential privacy, identity privacy, location privacy, IoT data privacy, privacy metrics, private federated learning, Multimedia security and privacy, cloud privacy, and forward privacy. C2: Security Data aggregation, rule engine, certificateless encryption, searchable encryption, authentication, multifactor authentication, data security, secure transmission, encryption, access control, confidentiality, integrity, decryption, and blockchain. C3: Industry 4.0 and Cyber physical systems Cyber Physical Systems, Industrial Cyber Physical Systems, Industry 4.0, and autonomous systems. C4: Cloud computing Cloud, edge, and fog computing. C5: Internet of Things Internet of Things, Industrial Internet of Things, and distributed Internet of Things. C6: Data analysis cluster Data analysis, data analytics, data mining, big data, diagnosis, monitoring, simulation, alarm systems, intrusion detection, anomaly detection, unsupervised clustering, scenarios, and Product Lifecycle Management (PLM). C7: AI Cluster Artificial Intelligence, machine learning, dictionary learning, deep learning, deep models, and reinforcement learning. C8: Federated learning Federated learning, data federation, transfer learning, federated dictionary learning, federated deep learning, and contrastive learning. Table 5. Mapping the reviewed studies to the defined categories in Table 4 Studies (in order of review) C1 C2 C3 C4 C5 C6 C7 C8 S1: (Wu et al., 2023) ✔ ✔ S2: (Jiang et al., 2021a) ✔ ✔ ✔ S3: (Sadique et al., 2020) ✔ ✔ ✔ S4: (Jiang et al., 2021b) ✔ ✔ ✔ ✔ S5: (Faujdar and Kaur, 2023) ✔ ✔ ✔ ✔ S6: (Li et al., 2023a) ✔ ✔ ✔ ✔ S7: (Huang et al., 2022a) ✔ ✔ ✔ S8: (Zainudin et al., 2022) ✔ ✔ S9: (Hinojosa-Palafox et al., 2021) ✔ ✔ S10: (Tajanpure and Muddana, 2023) ✔ ✔ S11: (Yang et al., 2020) ✔ S12: (Paul and Roslin, 2023) ✔ S13: (O’Donovan et al., 2019) ✔ ✔ ✔ ✔ S14: (Xu et al., 2021) ✔ ✔ ✔ S15: (Yang et al., 2022a) ✔ ✔ ✔ ✔ S16: (Pivoto et al., 2021) ✔ ✔ ✔ S17: (Zainudin et al., 2023) ✔ ✔ ✔ S18: (Shi et al., 2023) ✔ ✔ S19: (Zhang et al., 2020) ✔ ✔ S20: (Abdel-Basset et al., 2021) ✔ ✔ ✔ ✔ ✔ S21: (Das et al., 2022) ✔ ✔ S22: (Shrestha et al., 2024) ✔ ✔ S23: (Zhang et al., 2021) ✔ ✔ ✔ S24: (Cao et al., 2020) ✔ ✔ ✔ S25: (Huang et al., 2022b) ✔ ✔ S26: (Bokrantz et al., 2017) ✔ ✔ S27: (Ahmadi and Salehfar, 2022) ✔ ✔ ✔ S28: (Yang et al., 2022b) ✔ ✔ ✔ S29: (Kumar et al., 2021) ✔ S30: (Venkatesan et al., 2022) ✔ ✔ ✔ ✔ S31: (Li et al., 2023b) ✔ ✔ ✔ ✔ S32: (Li et al., 2020) ✔ ✔ ✔ ✔ ✔ S33: (Sharma et al., 2023) ✔ ✔ S34: (Milicic et al., 2017) ✔ ✔ ✔ Literature Survey Data privacy is a growing concern for the industry. For instance, Faujdar and Kaur (2023) highlight the threats to data privacy from both technological and management perspectives. Jiang et al. (2021b) discuss the risks of industrial data being used in warehouses and automation, including the potential for leakage during transmission through, e.g., Wi-Fi or other wireless media. Li et al. (2023a) emphasize the issue of commercial collaborations due to the risks of exposing users’ private data, and that faulty data might still include sensitive records. Tajanpure and Muddana (2023) emphasize the risks of unrestricted access to individuals’ information records, while Xu et al. (2021) discuss the potential of losing competitive advantages due to the use of users’ data for optimization and training algorithms. Li et al. (2023b) examine the latter issue closely, as privacy regulations restrict the exchange of industrial data between entities, which itself results in issues related to less optimization and efficiency. Yang et al. (2022a) extend the discussion to include multimedia data privacy risks, since multimedia data is of high criticality to privacy and is used in deployments, such as in IoT. Additionally, they raise concerns about the level of privacy, feasibility, heterogeneity of systems, computational overhead, transparency, and governance. Finally, Ahmadi and Salehfar (2022), Venkatesan et al. (2022), and Li et al. (2020) discuss the potential of cloud services that access sensitive personal and industrial data, e.g., personal, smart-home, energy, etc. Solutions related to data privacy have been closely examined and given special consideration due to their impacts. For instance, Wu et al. (2023) and Xu et al. (2021) point out that differential privacy techniques – though still in their early stages – can provide better privacy protection than current solutions that rely mostly on encryption. This is because differential privacy is computationally efficient and depends on adding noise into datasets according to a predefined privacy budget. This technique is also suggested by Jiang et al. (2021a), Tajanpure and Muddana (2023), Jiang et al. (2021b), and Yang et al. (2022a). In (Wu et al., 2023), a Software-Defined Network (SDN) algorithm based on centralized differential privacy with a trusted third party is proposed and proven to provide strong industrial data privacy protection and high availability. Jiang et al. (2021a) emphasize attack types and how they change the requirements for industrial data, and propose the use of a hybrid differential privacy combined with adaptive gradient compression of data. Tajanpure and Muddana (2023) suggest the use of differential privacy techniques with a convolution-based privacy-preserving algorithm that transforms data into lower dimensions, and thus preserves its privacy. Xu et al. (2021) highlight the importance of balancing protection and accuracy by controlling the privacy budget and propose the use of a privacy-preserviNg InCEntive (NICE) mechanism based on differential privacy. Similarly, Venkatesan et al. (2022) raise concerns about increased computational complexity, costs, performance issues, recommending the use of forward privacy property due to its efficiency and cost-effectiveness. On the other hand, Jiang et al. (2021b) state that industrial data differs from social data in aspects such as volume and required computational power, which could hinder the application of differential privacy techniques. They further highlight that network security, data value, and interconnection protocols are the main considerations for maintaining industrial data privacy. Sadique et al. (2020) observe that no existing study has identified all the points where end-user and industrial data privacy risks exist. Then, they propose a privacy enhancing framework focusing on security, communication, and gateway security; a layered approach; role-based authentication and access rules; restricted data sharing; intelligent privacy enhancing services; raising awareness; and law enforcement. Faujdar and Kaur (2023) propose a similar framework, and like Li et al. (2023a), suggest using decentralized identifiers to enhance privacy. Ahmadi and Salehfar (2022) make similar suggestions, emphasizing encryption, anonymization, and the lifecycle of privacy-preserving systems, i.e., design, verification, implementation, and deployment. Security is a fundamental requirement for protecting data and maintaining industrial data privacy, which is why security and privacy are often discussed together. For instance, Paul and Roslin (2023) discuss Wireless Sensor Networks (WSNs) and methods for protecting the security and privacy of aggregated data. Yang et al. (2022a) focus on IoT and the security and privacy of multimedia data. Shi et al. (2023) look at industrial data flow scenarios, examining how rule engines and access control mechanisms can be used to improve security and privacy. Zhang et al. (2020) and Venkatesan et al. (2022) investigate searchable encryption techniques, targeting the security and privacy of stored data. Li et al. (2020) similarly investigate homomorphic encryption, highlighting IIoT demands and focusing on improving the security and privacy of stored data. Das et al. (2022) look at multifactor authentication and assess the security and privacy of shared data transmitted over open channels. Finally, Cao et al. (2020) overview edge computing and examine security and privacy challenges and associated risks. As noticed, the reviewed studies highlighted various security measures, e.g., confidentiality, integrity, availability, utility, authentication, non-repudiation, and access control, which when effectively implemented help protect data privacy. Moreover, some studies (Zhang et al., 2020; Das et al., 2022; Venkatesan et al., 2022) referred to privacy as a key security measure, sometimes using the terms of privacy and confidentiality interchangeably. For example, Paul and Roslin (2023) discuss a homomorphic encoding method that enables the aggregation of encoded data, thus reducing the communication cost and guaranteeing accuracy, integrity, and authentication. Additionally, they explore certificateless signcryption for lightweight devices and RSA encryption to secure the order information of multifaceted data, thus enhancing the privacy of aggregated data. Yang et al. (2022a) identify cryptography, data hiding, chaos-based, and clustering-based schemes among the means employed to protect security and privacy. They also advise against the use of encryption techniques such as RSA, DES, and AES with multimedia data. Besides encryption, Shi et al. (2023) emphasize access control, identity verification, authentication, authorization, and intrusion detection for establishing a control mechanism to limit data access and identify anonymization as a key method to protect sensitive data. Zhang et al. (2020) stress the role of encryption in protecting industrial data privacy, proposing a Verifiable Certificateless Public Key Searchable Encryption (VCLPKSE) scheme capable of resisting two types of adversaries related to forging and replacing master encryption keys while authenticating the data owner’s identity. Das et al. (2022) emphasize lightweight security schemes to authenticate communicating devices, suggesting a multi-factor authentication scheme with low computation and communication costs, utilizing smart cards, passwords, Physical Unclonable Function (PUF), and fuzzy extractor. Cao et al. (2020) consider searchable encryption and discuss an attribute-based authentication and authorization framework that uses attributes and distributed certificates to replace traditional P2P networks’ public key certificates and access control authentication mechanisms to protect privacy. Furthermore, the study explores technologies for unified, cross-domain, and handover authentication, explicitly stating that access control is the key technology for ensuring system security and protecting user privacy. Ahmadi and Salehfar (2022), among others, highlight cryptography, anonymization, and authentication. Venkatesan et al. (2022) present a Lightweight Searchable Encryption and Delegation (LSED) Scheme for secure data storage and retrieval. Li et al. (2020) propose a Robust Cramer Shoup Delay Optimized Fully Homomorphic (RCS-DOFH) encryption technique, surpassing conventional encryption methods and ensuring secure data transmission. Lastly, blockchain technology has emerged as an innovative and modern technique for securing data and protecting privacy (Jiang et al., 2021b; Yang et al., 2022a; Abdel-Basset et al., 2021; Yang et al., 2022b; Sharma et al., 2023), with its ability to overcome the tradeoff between productivity and privacy, thus can be used for storing information, establishing trust, and protecting transactions. Finally, concerning risks, Paul and Roslin (2023) mention that the two main security challenges are confidentiality and integrity, while Zhang et al. (2020) stress these factors industrial data is handled externally. Yang et al. (2022a) emphasize the importance of determining the level of security and privacy and their associated computational overhead costs. In (Shi et al., 2023), stability, scalability, and real-time performance are among the issues highlighted. Cao et al. (2020) emphasize that the need for lightweight data encryption, sharing mechanisms, and multi-source heterogeneous data propagation control is among the challenges facing security and privacy. The issue of lightweight encryption is also emphasized by Ahmadi and Salehfar (2022). Yang et al. (2022b) overview efficiency issues related to security. Lastly, Sharma et al. (2023) highlight blockchain limitations, including insufficient computation resources and data breach risks, which may lead to information misuse and privacy threats. Industry 4.0 (I4.0) and the emergence of Cyber-Physical Systems (CPS) play a critical role in transforming the industrial landscape and changing the perception of industrial data. For instance, Hinojosa-Palafox et al. (2021) highlight a previous architecture that collects and integrates industrial data and Manufacturing Information Systems (MIS). They then propose an analytics architecture for Industrial CPS (ICPS) and discuss industrial process challenges. O’Donovan et al. (2019) compare I4.0 computing models and mention that ICPS are the primary enabler for I4.0, combining legacy industrial and control engineering with emerging technology paradigms. Similarly, Pivoto et al. (2021) survey the architecture of CPS for IoT applications in I4.0, emphasizing their characteristics, objectives, advantages, and contributions to I4.0. Zainudin et al. (2023) studies a federated learning Intrusion Detection (ID) and classification framework for SDN, highlighting the critical role of ICPS in storing actual network information, including personal data of manufacturing firms. Abdel-Basset et al. (2021) address the same topic, discussing the convergence of I4.0 with other technologies, ICPS, and the tradeoff between accuracy and privacy, and propose a federated learning threat hunting model suitable for ICPS owners due to its efficiency, awareness of heterogeneity, and privacy-preserving characteristics. Bokrantz et al. (2017) cover I4.0 from the angle of maintenance in digitalized manufacturing and highlight resistance to changing traditional systems, along with challenges of security, privacy, liability, and data ownership. Kumar et al. (2021) highlight the barriers in endorsing I4.0, which affect mitigation strategies, and emphasize the sensitivity of industrial data privacy concerning confidential organizational details. They also emphasize the I4.0 resistance issue and discuss data management, training, skills, and legal policies. Lastly, Milicic et al. (2017) address autonomous systems and Product Lifecyle Management (PLM) and present an ontology for protecting privacy of data. Concerning industrial data privacy, several topics have been discussed in the literature. Hinojosa-Palafox et al. (2021) highlight ICPS analytics privacy issue, concerns related to handling sensitive data without taking appropriate measures, and the potential for external servers to access confidential information. O’Donovan et al. (2019) stress that I4.0 can ensure appropriate levels of industrial data privacy through strict governance and firewall policies on automation and control networks. Yet, the real-time nature of industrial data and the need to sometimes send it outside the premises still pose a significant challenge. Pivoto et al. (2021) highlight the increased number of processing devices and systems, which can potentially pose privacy threats. They also emphasize the challenge of protecting user and industrial data privacy and examine the privacy of CPS. Zainudin et al. (2023) focus on the issue of sharing original data and the use of public networks. Abdel-Basset et al. (2021) concern the transmission of large volumes of data and associated privacy threats, emphasizing the criticality of maintaining industrial data privacy. In (Bokrantz et al., 2017), the dilemma of sharing data throughout supply chain for collaboration and related privacy concerns are highlighted. Kumar et al. (2021) discuss the importance of I4.0 data privacy and raise concerns that strict privacy could undermine the openness of various integrated digital infrastructures. They also mention that data privacy and theft issues are among the cyberattacks that could disrupt vital infrastructures, emphasizing that legal regulations and technical measures, such as hardware encryption and secure transit data networks, are essential. Lastly, Li et al. (2023b) discuss the use of data aggregation to address industrial data security and privacy concerns. Other concerns and challenges related to industrial data in I4.0 and CPS have also been identified. For instance, Hinojosa-Palafox et al. (2021) highlight the criticality of real-time industrial data processing. O’Donovan et al. (2019) emphasize performance, reliability, interoperability, and resilience. Bokrantz et al. (2017) address the lack of understanding regarding specific issues faced by maintenance organizations and emphasize the need for developing long-term strategy scenarios for realizing digitalized manufacturing. Additionally, they discuss socio-ethical aspects such as competence, training, education, and knowledge sharing. Kumar et al. (2021) highlight challenges facing the manufacturing industry, including high investments, unclear benefits, a lack of skilled workforce, resistance to change, inadequate infrastructure, data management, and legal policy constraints. Li et al. (2023b) further emphasize the challenges of fault prediction and decision-making. Lastly, Milicic et al. (2017) point out that a fully automated system does not yet exist. Cloud Computing Cloud computing \RL plays a major role in the realization and advancement of I4.0. As highlighted in the categorization section, cloud computing is a broad term encompassing three models, i.e., cloud, fog, and edge, which operate hierarchically (O’Donovan et al., 2019; Cao et al., 2020). These models are used to deliver on-demand resources, such as computing power, storage, and service access to end users within various contexts, criteria, and limitations. That being said, cloud computing – characterized by its processing power – operates within the core and data centers. Fog computing is for data processing, management, and transmission, while edge computing with its low latency, high performance, and consistency focuses on distribution and executing tasks closer to end users. Due to its significance for the modern industry, the literature has paid considerable attention to cloud computing and associated privacy issues. For instance, Jiang et al. (2021a) emphasize the need for industrial data privacy and raise concerns about emerging attack types and the requirements these impose on industrial edge computing to address the risks of industrial data leakage. Sadique et al. (2020) and Faujdar and Kaur (2023) overview IoT privacy risks, focusing on cloud devices and edge computing. O’Donovan et al. (2019) compare fog and cloud computing cyber-physical interfaces in terms of latency and reliability, highlighting the privacy challenges associated with controlling data transmitted outside the network boundaries. They also mention that fog computing can help mitigate this issue. Pivoto et al. (2021) focus more on CPS and industrial systems, highlighting the need to implement cloud computing broadly while addressing I4.0 requirements and the privacy challenges associated with large data flows. Cao et al. (2020) overview edge computing comprehensively, stating that traditional cloud computing is no longer sufficient to meet current needs. They suggest that edge computing can address these limitations by processing data locally, thereby mitigating security and privacy issues related to data leakage, data loss, and cyberattacks. Furthermore, they outline a list of security and privacy challenges facing edge computing, highlighting data outsourcing and trust issues, the need for lightweight data encryption and data-sharing control schemes, and the importance of combining traditional privacy protection methods with edge processing. Ahmadi and Salehfar (2022) discuss the topic of privacy-preserving cloud computing along with its associated ecosystem (anonymization, authentication, access control, cryptography, and watermarking) and lifecycle (design, verification, implementation, and deployment). They emphasize the importance of privacy for cloud computing for ensuring the integrity, accuracy, and accessibility of the stored data. Furthermore, they identify privacy as the most critical security aspect of cloud computing and list the types of industrial data stored in the cloud. Yang et al. (2022b), while discussing edge computing and blockchain, highlight concerns about industrial enterprises uploading their production and related data to the cloud. Venkatesan et al. (2022) focus on cloud privacy issues resulting from IIoT systems and along with Li et al. (2020) emphasize the potential of untrusted cloud services and the need for additional data collection measures when data is outsourced. Finally, Sharma et al. (2023) highlight privacy conflicts, efficiency, and computation complexities associated with cloud computing. They explicitly state that no single cloud solution can provide optimum privacy, therefore several solutions should be considered simultaneously. On the other hand, many solutions have been proposed to address cloud privacy issues. For instance, Jiang et al. (2021a) propose a data privacy scheme for industrial edge computing utilizing federated learning, combining different privacy models along with differential privacy. Sadique et al. (2020) propose a data privacy framework and advocate for cloud layer data privacy enforcement, cloud-limited data sharing, and the implementation of edge intelligence to enhance data privacy. Faujdar and Kaur (2023) add enforcement of rules, policies, and laws as well as user awareness and legal education to the equation. Cao et al. (2020) suggest that processing data nearby in edge computing provides better privacy protection, and that computing offloading, mobility management, traffic offloading, and network control technologies can help with edge privacy. Moreover, they along with Venkatesan et al. (2022) emphasize the importance of searchable encryption as one of the key solutions for protecting privacy and data stored in the cloud without suffering from computational complexity and increased costs. Ahmadi and Salehfar (2022) highlight the need for categorizing data according to sensitivity, thus preserving it while reducing costs. They also emphasize the need for integrity-by-design data sharing systems, lightweight layered privacy architecture, and data auditing. Then along with Li et al. (2020), they mention the solution of encrypting data before uploading it to the cloud and performing operations on encrypted data rather than raw data. Yang et al. (2022b) suggest edge computing to improve the efficiency of parallel model training, thus ensuring the privacy of industrial data. Lastly, Sharma et al. (2023) present a cloud-assisted, secured protocol based on Elliptical Curve Cryptography (ECC) to enhance privacy through optimized key generation, encryption, and decryptions times. Similarly, Internet of Things (IoT) technology plays a vital role in the industrialization and realization of I4.0 objectives. This technology, which includes paradigms such as the Industrial Internet of Things (IIoT), the Internet of Everything (IoE), etc., incorporates connected sensor devices and data transmission means, enabling real-time monitoring, smart operations, event sensing, data analysis, self-decision, and process optimization (Jiang et al., 2021b; Das et al., 2022). However, with the proliferation of interconnected smart devices and the substantial amounts of data being collected, careful consideration of both security and privacy is required (Jiang et al., 2021b; Pivoto et al., 2021; Abdel-Basset et al., 2021; Das et al., 2022; Zhang et al., 2021; Cao et al., 2020). Sadique et al. (2020) – for instance – highlight the issue of IoT data privacy and identify the areas where end-user and industry risks exist. Specifically, end devices, gateways, mobile devices, and communication channels are among the places where IoT risks are present. Jiang et al. ( 2021b) suggest that IIoT is an extension of cloud and edge computing; hence, the issue of industrial data privacy is very important, especially considering the increase in data volumes. Faujdar and Kaur (2023) address the issue of IoT sharing private users’ data. This is also raised by Xu et al. (2021), where clients’ data carried by IIoT applications are revealed during machine learning model training. Yang et al. (2022a) discuss multimedia data and the importance of security and privacy in IoT systems, examining how these factors can impact trustworthiness. Additionally, they classify multimedia into five categories and security and privacy into three levels, thus allowing for the implementation of effective measures. Zhang et al. (2020) and Venkatesan et al. (2022) explore the storage of IIoT data in the cloud, highlighting the importance of security and privacy research. On a different topic, Abdel-Basset et al. (2021) note that IoT and I4.0 have increased the vulnerability of ICPSs, while Das et al. (2022) highlight that IIoT has introduced new risks, making devices more susceptible to various attacks, such as cloning, impersonation, man-in-the-middle, and physical attacks. Then, Cao et al. (2020) discuss the risks of IoT data leakage during transfer and the need for new data sharing and governance requirements for interconnected IoT devices to ensure effective privacy. On the other hand, several approaches have been suggested to address the points outlined above. In (Sadique et al., 2020), a framework for data privacy enhancement is suggested, emphasizing protection at IoT gateways. Jiang et al. (2021b) distinguish between social and industrial data and provide differential privacy applications for IIoT. Xu et al. (2021) suggest several solutions, including encryption, secure multiparty computation, and differential privacy as a lightweight tool for data privacy. Yang et al. (2022a) identify several security and privacy requirements for IoT that can be addressed through schemes like cryptography, data hiding, chaos-based methods, and blockchain. Zhang et al. (2020) focus on encryption solutions, highlighting VCLPKSE as a lightweight encryption scheme for protecting IIoT data before uploading ciphertext to cloud servers. This approach addresses trustworthiness issues in the IIoT environment by authenticating the data owner’s identity and resisting two types of adversaries, thereby enhancing security and privacy measures. Abdel-Basset et al. (2021) address privacy, eavesdropping, and data leakage issues by proposing a federated learning approach to add intelligence to the edge layers of IoT networks (Zhang et al., 2021). In (Das et al., 2022), a Physical Unclonable Function (PUF) with a Fuzzy Extractor (FE) is proposed as a two-factor authentication scheme for protecting IoT devices. Cao et al. (2020) suggest the previously mentioned searchable encryption solution, focusing on identity authentication and access control to ensure systems’ security and data privacy. Venkatesan et al. (2022) emphasize the same solution and propose a Lightweight Searchable Encryption and Delegation (LSED) methodology with forward privacy to enable secure data storage and retrieval in IIoT-cloud systems. Finally, Li et al. (2020) provide a secure data transmission mechanism for IIoT with a privacy-preserving hash-based deep learning method in separate encryption and decryption to protect privacy and address latency and increased computational costs. Data analysis plays a pivotal role in modern industrialization, serving as a key enabler for advancements through providing insights and directions for improvements, optimization, and efficiency. The literature has accordingly paid considerable attention to this and related topics. For instance, Li et al. (2023a) emphasize the benefits of training models collaboratively between industries; however, privacy, permissions, and data sharing emerge as significant risks and critical issues. Similarly, Huang et al. (2022a) address the same topic, highlighting the challenges facing centralized monitoring due to not sharing raw data externally, though privacy and leakage threats persist. Hinojosa-Palafox et al. (2021) discuss the architecture design of industrial data analytics concerning big data and ICPS, emphasizing privacy concerns and the handling of sensitive data. Tajanpure and Muddana (2023) address data mining applications and associated personal information risks; while they focus on statistically useful patterns, these applications still pose a threat of unrestricted access to records. Additionally, the authors emphasize the importance of sharing analytics privacy in I4.0 and the need for privacy-preserving techniques for real-time mining. Yang et al. (2020) highlight the conflict between privacy policies and sharing data analysis and new results, emphasizing the need for a large-scale simulated alarm system to create event data for testing new methods. Zainudin et al. (2023) discuss the privacy issues associated with centralized deep learning-based Intrusion Detection System (IDS) as well as the communication overhead. On a different note, Shi et al. (2023) address the topic by highlighting the gap and the need for a trusted and controllable data management platform and ecosystem to handle industrial data issues. Additionally, they discuss challenges in data flow based on user behavior analysis results, such as security, privacy, and performance. Shrestha et al. (2024) discuss knowledge extraction techniques and their potential for data misuse, manipulation, or privacy leakage. Moreover, they state that privacy is not well-considered in the design of smart grid systems due to the assumption that systems could be isolated. Finally, Milicic et al. (2017) suggest that independent data mining systems cannot be fully automated. On the solutions side, Li et al. (2023a) stress that traditional centralized data aggregation approaches should be avoided completely to protect privacy, opting for federated learning and federated data for collaboration when needed. In (Huang et al., 2022a), distributed K-Singular Value Decomposition (K-SVD) method is suggested for centralized data collection, as it can perform monitoring tasks without sharing local data between nodes, thus preserving privacy. Hinojosa-Palafox et al. (2021) propose an architecture for collecting and integrating industrial data from IIoT and MIS, enabling subsequent industrial analytics. In (Tajanpure and Muddana, 2023), a convolution-based privacy-preserving algorithm that transforms data into lower dimensions, thus preserving its privacy while mining, is proposed. Such an algorithm benefits from better accuracy, data utility, and performance. In (Zainudin et al., 2023), the issue with deep learning-based IDS systems is addressed through a low-complexity federated learning-based IDS combined with a classification framework for SDN-based ICPS. This solution is becoming popular for industrial applications due to its effectiveness, privacy features, and low communication overhead. In (Shi et al., 2023), customizing and configuring strict, refined rule engines, assessing user access behavior before, during, and after data circulation, and establishing data access control and identity authentication mechanisms, are proposed. Bokrantz et al. (2017) suggest adjusting the legal framework to manage ongoing issues, including the privacy of industrial data, thus making data systems more efficient and socially sustainable. Finally, Milicic et al. (2017) present an ontology combined with PLM to protect industrial data privacy. Artificial Intelligence (AI) technologies are currently driving the industry forward through their advanced human-like intelligence and capability for effective and efficient analysis and decision-making. As a result, significant focus has been placed on this field and related areas, particularly due to the privacy concerns they may raise. For instance, Wu et al. (2023) examine Deep Learning, proposing an algorithm that provides strong industrial data privacy protection and high availability. Likewise, Li et al. (2023a) explore deep learning and emphasize collaboration for training powerful models, noting that centralized data aggregation model training is not preferred in real scenarios. Huang et al. (2022a) discuss dictionary learning methods and the privacy risks associated with industrial raw data, along with the increased risks from centralized data collection methods. O’Donovan et al. (2019) highlight the role of AI in enabling I4.0. In (Abdel-Basset et al., 2021), topics such as threat hunting and ICPS are covered, along with data federation and deep learning, highlighting the issues of training deep learning models centrally and the threats posed by transmitting data to other nodes. The study also underscores the challenges of developing distributed deep learning due to resources constraints and privacy matters, including eavesdropping and data leakage. Zhang et al. (2021) and Huang et al. (2022b) also examine deep learning training models, emphasizing the privacy concerns associated with traditional centralized training methods. In (Yang et al., 2022b), reinforcement learning is discussed in the context of edge computing, along with the role of parallel reinforcement learning for collaborative resource scheduling. Li et al. (2023b) highlight the role of AI in fault prediction, traffic analysis, and decision-making, discussing the regulatory challenges of transferring and exchanging industrial data between entities, which can limit the accuracy of AI models. Li et al. (2020) focus on encryption methods in IIoT and cloud computing in relation to AI. Lastly, Milicic et al. (2017) focus on autonomous systems for PLM and discuss the use of data mining and AI in manufacturing systems. Several solutions have been proposed in the literature to address some of the identified concerns. For example, Wu et al. (2023) suggest privacy protection through the use of synthetic data semantics and centralized differential privacy models. Similarly, Jiang et al. (2021b) propose generating noisy data to protect privacy. Faujdar and Kaur (2023) focus on conceptual aspects of privacy protection, such as enhancing security and awareness, as well as enforcing rules and policies. Li et al. (2023a) explicitly state that local data should not be shared for centralized AI model training to avoid privacy concerns. Huang et al. (2022a) suggest K-SVD dictionary learning method for monitoring without sharing local data between nodes, noting that a third party cannot infer the original data if it obtains the dictionary model, due to the over-completeness of the model, thus protecting industrial data privacy. O’Donovan et al. (2019) opt for decentralized intelligence for its benefits, including near real-time performance, privacy, and the openness and interoperability of systems. Abdel-Basset et al. (2021), Zhang et al. (2021), and Huang et al. (2022b) highlight collaborative learning and training models, for their efficiency addressing privacy issues. Li et al. (2023b) suggest data aggregation for training models following the centralized learning, though this method might compromise the security and privacy of industrial data. Finally, Li et al. (2020) recommend the use of AI in the form of convolutional neural networks to enhance privacy and present two deep learning schemes based on asynchronous patterns to preserve secrecy. They also note that combining deep learning with cryptographic methods offers several advantages, such as reducing computational costs and enhancing privacy preservation. Finally, although federated learning is a subset of data analysis and AI, it is highlighted separately in this section as it received special attention in the reviewed literature for its unique features, particularly decentralization and enhanced data privacy. For instance, Jiang et al. (2021a) suggest federated learning for industrial edge computing and highlight the need for emerging technologies to pay more attention to industrial data privacy. Li et al. (2023a) discuss federated transfer learning and emphasize the risks associated with centralized training. Huang et al. (2022a) examine federated dictionary learning in the field of monitoring, shedding light on economic and industrial risks of centralized data collection and emphasizing the risks associated with transmitting model parameters, hence the need for robust data privacy requirements in federated learning applications. Zainudin et al. (2022) discuss Distributed Denial of Service (DDoS) attacks in SDN-enabled IIoT networks and the capabilities of federated learning in building resilience against such attacks. Yang et al. (2022a) address the challenges facing data security and privacy in IoT, highlighting federated learning among other protection schemes such as cryptography, data hiding, chaos-based methods, blockchain, and clustering. Zainudin et al. (2023) discuss a federated learning IDS technique, highlighting privacy concerns and communication overhead associated with the use of centralized deep learning-based IDSs. Abdel-Basset et al. (2021) focus on federated threat hunting and discuss the limitations of distributed deep learning, emphasizing how federated learning addresses these issues. Additionally, they highlight the risks posed by central authorities, including their potential to provoke federated learning models if not fully trustworthy in managing model training. Shrestha et al. (2024) address anomaly detection along with the security and privacy concerns arising from poorly shared local models. Zhang et al. (2021) discuss federated learning techniques for distributed IIoT systems, emphasizing privacy concerns related to massive raw data from IIoT devices and highlighting the benefits of local data processing and model training. Huang et al. (2022b) discuss federated domain adaptation, highlighting the heterogeneity of local data, mutual information silos, and associated privacy risks. Finally, Li et al. (2023b) emphasize the use of federated learning to address the challenges of data silos and fragmented training data, pointing out the vulnerability of federated learning to interference and Byzantine attacks from aggregators and participants. Regarding contributions and proposed solutions, several were identified. For instance, Jiang et al. (2021a) propose the use of federated edge learning based on differential privacy and adaptive compression for industrial data processing. Li et al. (2023a) suggest the inclusion of synthetic data to protect data sources. Zainudin et al. (2022) emphasize decentralization through federated learning and the transmission of training parameters to an aggregation server, thus overcoming machine learning privacy concerns. Additionally, the authors mention the successful implementation of a federated learning-based IDS, also applied to DDoS attack classification. Xu et al. (2021) propose the use of NICE privacy mechanism and Stackelberg games for federated learning, thus offering flexibility, control over one’s own data, and preventing data and privacy leaks. Zainudin et al. (2023) suggest combining federated learning with ICPS to transform IIoT applications into significant industrial domains where privacy concerns are effectively addressed. They also emphasize federated learning-based IDS systems for their low complexity, computational efficiency, and effectiveness in preserving privacy. In (Abdel-Basset et al., 2021), a novel federated threat hunting approach is presented, providing threat intelligence solution suitable for all ICPS owners due to its efficiency, awareness of heterogeneity, and privacy-preserving characteristics against actors capturing network data. Shrestha et al. (2024) propose using federated learning for anomaly detection in smart grid systems with artificial neural networks, such as Long Short-Term Memory (LSTM) autoencoders, along with homomorphic encryption to ensure privacy and security throughout the model training process. Zhang et al. (2021) introduce a three-layer architecture of device-edge-cloud to support federated learning and optimize distributed IIoT networks by reducing backbone network traffic and parameter transmission. Huang et al. (2022b) propose an effective federated multi-source domain adaptation algorithm based on knowledge distillation and contrastive learning to train high accuracy models locally while protecting data privacy. Lastly, Li et al. (2023b) suggest a robust privacy-preserving Byzantine-based federated learning scheme (PBFL) that works for the majority or participants and aggregators, to protect privacy and leverage clustering. Synthesis and Discussion In this section we synthesize the findings and analyses by answering two key questions: “What are the implications?” and “What have we learnt and what should be done next?” (Campbell, 2012; McMahan and McFarland, 2021). First, industrial systems are inherently complex and diverse, spanning various sectors and industries with distinct concepts, frameworks, processes, and devices. To understand these systems and their characteristics, two approaches could be used: examining them either individually within their specific contexts or collectively in abstraction (Colin and His, 1997). While the former approach provides detailed insights of each system, the latter is more practical, given the continuous growth and evolution of industrial systems. Therefore, the collective approach is adopted in this work. Following this, our analysis highlights that modern industrial systems share common features, such as integration, connectivity, complexity, automation, distribution, coordination, specialization, and data-driven processes, with emphasis on optimization, efficiency, reliability, scalability, and resilience. While these features are widely shared, their prominence varies depending on the system and its scale of operation. For instance, Industrial Analytics and AI systems prioritize data-driven insights, while Safety Instrumented and Industrial Cybersecurity systems focus on reliability, and Industrial Communication Networks emphasize connectivity, interoperability, and scalability. This diversity in focus highlights the need for tailored approaches to address system-specific challenges while leveraging shared foundations. Second, industrial data is central to the capabilities and value of industrial systems, distinguishing them from small-scale, consumer-oriented, and traditional systems. For example, the ability of these systems to optimize operations, manage resources, and maintain quality standards relies on the effective utilization of industrial data. Similarly, the value of industrial data varies across sectors and systems, depending on how insights and patterns are leveraged, as shown in Table 2. However, this value also introduces significant risks, particularly regarding data privacy and protection, given the potential negative impacts of breaches, misuse, and data manipulation. As our preliminary survey indicates, although awareness of these risks is growing, privacy issues have primarily been addressed for individuals and organizations. Only recently has the topic of industrial data privacy gained more attention in research. This delay highlights the need for further work to develop a comprehensive understanding of industrial data and its contributing factors to protect it while preserving its benefits. Third, the reviewed literature identifies eight major, distinct yet overlapping themes related to industrial data privacy, which are grouped into three areas. The first – privacy and security themes – addresses core concepts and measures for protecting data. The second – Industry 4.0 and Cyber-physical systems, Internet of things, and cloud computing themes – focuses on technological advancements that support manufacturing and drive the convergence of physical and digital systems. The third, on the cutting edge – data analysis cluster, Artificial intelligence, and Federated Learning themes – concerns deriving value and insights from data. These themes highlight not only different aspects of industrial systems and data but also the interdisciplinary nature of this field, alongside the challenges of protecting data privacy across these domains. Key challenges include data leakage, eavesdropping, and the exposure of private and sensitive records during collaborations, outsourcing, model learning, and analytics. Additionally, inadequate policies and legal regulations, real-time performance limitations, scalability issues, the increasing number of connected devices, device trustworthiness, and growing data volumes further complicate industrial data protection. Moreover, many challenges are not purely technical but also operational, such as balancing efficiency with accuracy. This is evident in AI-driven systems, where increased privacy measures can degrade model accuracy and limit optimization and decision-making capabilities. Similarly, resource-intensive solutions often conflict with cost-effectiveness, which highlights the need for scalable and practical approaches. Fourth, a wide range of governance and technical solutions covering different aspects has been proposed. Governance efforts, such as raising awareness and enforcing privacy regulations, lay the foundation for protection industrial data. Nonetheless, technical solutions addressing both conceptual and application challenges are essential to support these efforts. Among the conceptual solutions are anonymization, data hiding, and chaos-based techniques to obscure data and its sources. Differential privacy adds noise to data while preserving its valuable insights. Adaptive gradient compression and convolution-based algorithms reduce data dimensionality. The forward privacy property ensures secure data storage and retrieval. A layered privacy approach safeguards data, decentralized identifiers enhance privacy, and intelligent services restrict data sharing. On the application side, solutions include implementing strict access control mechanisms, using homomorphic encryption for secure processing without decryption, employing searchable encryption to search encrypted data, replacing traditional P2P authentication with attribute-based and distributed certificates, applying certificateless signcryption for lightweight devices, utilizing elliptical curve cryptography for efficient key generation, and adopting clustering-based and blockchain solutions to ensure security, privacy and trust. Alongside these, edge computing has emerged as a promising approach to decentralize data processing, providing necessary computational resources while improving privacy by minimizing data transmission and exposure. When combined with computing offloading, mobility management, and network control technologies, privacy can be further enhanced. Similarly, federated learning offers advanced privacy protection by keeping data localized while enabling collaborative insights. Practical solutions in this domain include using federated learning combined with differential privacy for edge computing, performing data limitation, and implementing intelligence at the edge. Furthermore, deploying secure multiparty computation for lightweight implementations, working with aggregated data and decentralized intelligence, and transmitting model parameters instead of actual models and associated data can further improve privacy. The use of distributed K-SVD for data collection enables monitoring without sharing local data between nodes, and multi-source domain adaptation algorithms based on knowledge distillation and contrastive learning help train accurate models locally, further improving privacy. Fifth, despite these advancements and solutions, significant gaps and open challenges remain. These include the continuous development of industrial systems and the need to adopt privacy-by-design principles to embed privacy measures into the core architecture of future systems. Categorizing industrial data based on its value and sensitivity could provide cost-effective solutions while ensuring appropriate protection levels. Lastly, addressing workforce skill gaps and change management challenges, as well as developing scenarios that emphasize the value and benefits of privacy-centric manufacturing, are essential. Conclusions and Future work This study examines industrial systems, their roles, and their reliance on industrial data to perform critical operations and advanced functionalities. It then focuses on industrial data, emphasizing its value and the importance of protecting its privacy amidst the increasing risks facing manufacturing. As industries transform and advance, reliance on industrial systems and in turn industrial data continues to grow. While this transformation offers numerous benefits, it also introduces significant risks related to privacy and the protection of data. The findings highlight the complexity of the topic, and despite the existence of various governance and technical solutions, significant challenges remain. Accordingly, there is a strong need for adopting a multi layered approach and developing comprehensive, holistic solutions that consider industrial data privacy from the outset than relying on individual system-specific solutions that might require later adjustments. Modern privacy-native technologies, such as edge computing, differential privacy, and federated learning not only provide operational advantages, but also could enhance privacy protection, thus they should be considered more. Additionally, conceptual approaches such as privacy-by-design principles, cross-sector collaboration, and addressing awareness and skill gaps are equally critical. Future research related to this topic should focus on Privacy Enhancing Technologies (PET), adaptive privacy frameworks, operational constraints such as efficiency and computational challenges, and improving the interoperability of privacy solutions. References 1. Brennan, Louis, et al. ”Manufacturing in the world: where next?.” International Journal of Operations & Production Management 35.9 (2015): 1253-1274. 2. Agarwal, Nivedita, and Alexander Brem. ”Strategic business transformation through technology convergence: implications from General Electric’s industrial internet initiative.” International Journal of Technology Management 67.2-4 (2015): 196-214. 3. ElMaraghy, Hoda, et al. ”Evolution and future of manufacturing systems.” CIRP Annals 70.2 (2021): 635-658. 4. Lu, Yang. ”Industry 4.0: A survey on technologies, applications and open research issues.” Journal of industrial information integration 6 (2017): 1-10. 5. Zhong, Ray Y., et al. ”Intelligent manufacturing in the context of industry 4.0: a review.” Engineering 3.5 (2017): 616-630. 6. Bordeleau, Fanny-Eve, Elaine Mosconi, and Luis Antonio Santa-Eulalia. ”Business Intelligence in Industry 4.0: State of the art and research opportunities.” (2018). 7. Wang, David. ”Building value in a world of technological change: Data analytics and industry 4.0.” IEEE Engineering Management Review 46.1 (2018): 32-33. 8. Martínez, Patricia López, et al. ”A big data-centric architecture metamodel for Industry 4.0.” Future Generation Computer Systems 125 (2021): 263-284 9. Kayabay, Kerem, et al. ”Data science roadmapping: An architectural framework for facilitating transformation towards a data-driven organization.” Technological Forecasting and Social Change 174 (2022): 121264. 10. Raptis, Theofanis P., Andrea Passarella, and Marco Conti. ”Data management in industry 4.0: State of the art and open challenges.” IEEE Access 7 (2019): 97052-97093. 11. Corallo, Angelo, et al. ”Cybersecurity challenges for manufacturing systems 4.0: assessment of the business impact level.” IEEE Transactions on Engineering Management 70.11 (2021): 3745-3765. 12. Haleem, Abid, et al. ”Perspectives of cybersecurity for ameliorative Industry 4.0 era: A review-based framework.” Industrial Robot: the international journal of robotics research and application 49.3 (2022): 582-597. 13. Onik, Md Mehedi Hassan, K. I. M. Chul-Soo, and Y. A. N. G. Jinhong. ”Personal data privacy challenges of the fourth industrial revolution.” 2019 21st International Conference on Advanced Communication Technology (ICACT) . IEEE, 2019. 14. Waidner, Michael, and Michael Kasper. ”Security in industrie 4.0-challenges and solutions for the fourth industrial revolution.” 2016 Design, Automation & Test in Europe Conference & Exhibition (DATE) . IEEE, 2016. 15. Timan, Tjerk, and Zoltan Mann. ”Data protection in the era of artificial intelligence: trends, existing solutions and recommendations for privacy-preserving technologies.” The elements of big data value: Foundations of the research and innovation ecosystem . Cham: Springer International Publishing, 2021. 153-175. 16. Sun, Liyuan, Hongyun Zhang, and Chao Fang. ”Data security governance in the era of big data: status, challenges, and prospects.” Data Science and Management 2 (2021): 41-44. 17. Guan, Zhongqi. ”Difficulties and Solutions for Industrial Data Security and Compliance Governance.” International Conference on Cloud Computing . Cham: Springer Nature Switzerland, 2023. 18. Chevallier, Arnaud. Strategic thinking in complex problem solving . Oxford University Press, 2016. 19. The Royal Academy of Engineering. “Industrial Systems: capturing value through manufacturing”. The Royal Academy of Engineering , 2012. https://raeng.org.uk/media/yurbqhvx/industrial_systems_capturing_value_through_manufacturing.pdf (accessed 25 August 2024) 20. Jamil, Uzair, and Joshua M. Pearce. ”Energy Policy for Agrivoltaics in Alberta Canada.” Energies 16.1 (2022): 53. 21. Holler, Jan, et al. Internet of things . Academic Press, 2014. 22. Badiru, Adedeji, Abidemi Badiru, and Adetokunboh Badiru. Industrial project management: Concepts, tools, and techniques . CRC Press, 2007. 23. Luo, Changkun. ”Application Prospect of Data Mining Technology in Intelligent Manufacturing.” Academic Journal of Science and Technology 3.1 (2022): 41-43. 24. Urbani, Michele. ”Maintenance policies optimization in the Industry 4.0 paradigm.” (2021). 25. Veblen, Thorstein. The theory of business enterprise . New English Library, 2003. 26. Levine, Stephen H. ”Comparing products and production in ecological and industrial systems.” Journal of industrial ecology 7.2 (2003): 33-42. 27. Boons, Frank, Wouter Spekkink, and Yannis Mouzakitis. ”The dynamics of industrial symbiosis: a proposal for a conceptual framework based upon a comprehensive literature review.” Journal of cleaner production 19.9-10 (2011): 905-911. 28. Whitelock, Jeryl. ”Theories of internationalisation and their impact on market entry.” International marketing review 19.4 (2002): 342-347. 29. Riahi-Belkaoui, Ahmed. Behavioral management accounting . Bloomsbury Publishing USA, 2001. 30. Dai, Xuewu, and Zhiwei Gao. ”From model, signal to knowledge: A data-driven perspective of fault detection and diagnosis.” IEEE Transactions on Industrial Informatics 9.4 (2013): 2226-2238. 31. Qi, Saiyu, et al. ”Cpds: Enabling compressed and private data sharing for industrial Internet of Things over blockchain.” IEEE Transactions on Industrial Informatics 17.4 (2020): 2376-2387. 32. Huo, Cui Hua, and Li He Chai. ”Physical principles and simulations on the structural evolution of eco-industrial systems.” Journal of Cleaner Production 16.18 (2008): 1995-2005. 33. Bas, Gerben. ”Resilient Industrial Systems: A Complex System Perspective to Support Business Decisions.” (2017). 34. Macaulay, Tyson, and Bryan L. Singer. Cybersecurity for industrial control systems: SCADA, DCS, PLC, HMI, and SIS . CRC Press, 2011. 35. Bhardwaj, Akashdeep, et al. ”Capturing-the-invisible (CTI): Behavior-based attacks recognition in IoT-oriented industrial control systems.” IEEE access 8 (2020): 104956-104966. 36. Campilho, Raul DSG, and Francisco JG Silva. ”Industrial Process Improvement by Automation and Robotics.” Machines 11.11 (2023): 1011. 37. Javaid, Mohd, et al. ”Exploring impact and features of machine vision for progressive industry 4.0 culture.” Sensors International 3 (2022): 100132. 38. Campbell, Thomas, et al. ”Could 3D printing change the world.” Technologies, Potential, and Implications of Additive Manufacturing, Atlantic Council, Washington, DC 3.1 (2011): 18. 39. Liu, Dan, et al. ”Communication Quality Indicator System of Industrial Communication Network and Related Diagnostic Evaluation.” (2022). 40. Harmatos, János, and Markosz Maliosz. ”Architecture integration of 5G networks and time-sensitive networking with edge computing for smart manufacturing.” Electronics 10.24 (2021): 3085. 41. Gröger, Christoph. ”Industrial analytics–An overview.” it-Information Technology 64.1-2 (2022): 55-65. 42. Wang, Lihui. ”From intelligence science to intelligent manufacturing.” Engineering 5.4 (2019): 615-618. 43. Grewal, Dalvinder Singh. ”A critical conceptual analysis of definitions of artificial intelligence as applicable to computer engineering.” IOSR Journal of Computer Engineering 16.2 (2014): 9-13. 44. Perwej, Yusuf, et al. ”The internet of things (IoT) and its application domains.” International Journal of Computer Applications 975.8887 (2019): 182. 45. Helmiö, Petra. ”Open source in industrial internet of things: A systematic literature review.” (2017). 46. Liu, Yang, et al. ”Review on cyber-physical systems.” IEEE/CAA Journal of Automatica Sinica 4.1 (2017): 27-40. 47. D’Antonio, Gianluca, Joel Sauza Bedolla, and Paolo Chiabert. ”A novel methodology to integrate manufacturing execution systems with the lean manufacturing approach.” Procedia Manufacturing 11 (2017): 2243-2251. 48. Avramovic, B., and L. H. Fink. ”Energy management systems and control of FACTS.” International Journal of Electrical Power & Energy Systems 17.3 (1995): 195-198. 49. Paraschivescu, Andrei Octavian. ”The advantages of the process of integrating quality management system.” Economy Transdisciplinarity Cognition 19.2 (2016): 48. 50. International Organization for Standardization. ISO 9000: 2015 quality management systems-fundamentals and vocabulary . 2015. 51. Werbińska-Wojciechowska, Sylwia, and Klaudia Winiarska. ”Maintenance performance in the age of Industry 4.0: A bibliometric performance analysis and a systematic literature review.” Sensors 23.3 (2023): 1409. 52. Schwab, Wolfgang, and Mathieu Poujol. ”The state of industrial cybersecurity 2018.” Trend Study Kaspersky Reports 33 (2018). 53. Xu, Yin, et al. ”Incentive mechanism for differentially private federated learning in industrial Internet of Things.” IEEE Transactions on Industrial Informatics 18.10 (2021): 6927-6939. 54. Okoli, Chitu, and Kira Schabram. ”A guide to conducting a systematic literature review of information systems research.” (2015). 55. Rowe, Frantz. ”What literature review is not: diversity, boundaries and recommendations.” European Journal of Information Systems 23.3 (2014): 241-255. 56. Schryen, Guido. ”Writing qualitative is literature reviews—guidelines for synthesis, interpretation, and guidance of research.” Communications of the Association for Information Systems 37.1 (2015): 12. 57. Wu, Wenjia, Qi Qi, and Xiaosheng Yu. ”Deep learning-based data privacy protection in software-defined industrial networking.” Computers and Electrical Engineering 106 (2023): 108578. 58. Jiang, Bin, et al. ”Privacy-preserving federated learning for industrial edge computing via hybrid differential privacy and adaptive compression.” IEEE Transactions on Industrial Informatics 19.2 (2021a): 1136-1144. 59. Sadique, Kazi Masum, Rahim Rahmani, and Paul Johannesson. ”Enhancing data privacy in the Internet of Things (IoT) using edge computing.” International Conference on Computational Intelligence, Security and Internet of Things . Cham: Springer International Publishing, 2020. 60. Jiang, Bin, et al. ”Differential privacy for industrial internet of things: Opportunities, applications, and challenges.” IEEE Internet of Things Journal 8.13 (2021b): 10430-10451. 61. Faujdar, Pramod Kumar, and Gurmandeep Kaur. ”A Framework for Internet of Things Services that Enhances Privacy.” 2023 3rd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE) . IEEE, 2023. 62. Li, Xu, et al. ”Federated transfer learning in fault diagnosis under data privacy with target self-adaptation.” Journal of Manufacturing Systems 68 (2023): 523-535. 63. Huang, Keke, et al. ”A federated dictionary learning method for process monitoring with industrial applications.” IEEE Transactions on Artificial Intelligence 4.5 (2022a): 1017-1028. 64. Zainudin, Ahmad, et al. ”FedDDoS: An efficient federated learning-based DDoS attacks classification in SDN-enabled IIoT networks.” 2022 13th International Conference on Information and Communication Technology Convergence (ICTC) . IEEE, 2022. 65. Hinojosa-Palafox, Eduardo A., et al. ”An analytics environment architecture for industrial cyber-physical systems big data solutions.” Sensors 21.13 (2021): 4282. 66. Tajanpure, Rupali, and Akkalakshmi Muddana. ”Data analysis with performance and privacy enhanced classification.” Journal of Intelligent Systems 32.1 (2023): 20220215. 67. Yang, Guang, et al. ”Simulating industrial alarm systems by extending the public model of a vinyl acetate monomer process.” 2020 39th Chinese Control Conference (CCC) . IEEE, 2020. 68. Paul, Annie, and S. Emalda Roslin. ”A Brief Study on Security Preserved Data Aggregation Approaches in WSN s.” 2023 Advanced Computing and Communication Technologies for High Performance Applications (ACCTHPA) . IEEE, 2023. 69. O’Donovan, Peter, et al. ”A comparison of fog and cloud computing cyber-physical interfaces for Industry 4.0 real-time embedded machine learning engineering applications.” Computers in industry 110 (2019): 12-35. 70. Yang, Wencheng, et al. ”Multimedia security and privacy protection in the internet of things: research developments and challenges.” International Journal of Multimedia Intelligence and Security 4.1 (2022a): 20-46. 71. Pivoto, Diego GS, et al. ”Cyber-physical systems architectures for industrial internet of things applications in Industry 4.0: A literature review.” Journal of manufacturing systems 58 (2021): 176-192. 72. Zainudin, Ahmad, et al. ”Federated learning inspired low-complexity intrusion detection and classification technique for sdn-based industrial cps.” IEEE Transactions on Network and Service Management (2023). 73. Shi, Fangning, Zhongli Na, and Zhifeng Gao. ”Configurable Rule Engine for Industrial Data Flow Scenarios.” 2023 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB) . IEEE, 2023. 74. Zhang, Yulei, et al. ”VCLPKES: Verifiable certificateless public key searchable encryption scheme for industrial Internet of Things.” IEEE Access 8 (2020): 20849-20861. 75. Abdel-Basset, Mohamed, Hossam Hawash, and Karam Sallam. ”Federated threat-hunting approach for microservice-based industrial cyber-physical system.” IEEE Transactions on Industrial Informatics 18.3 (2021): 1905-1917. 76. Das, Ayan Kumar, et al. ”Macp: Multifactor authentication using physical unclonable function and fuzzy extractor based chebyshev polynomial for industrial internet of things devices.” Transactions on Emerging Telecommunications Technologies 33.11 (2022): e4581. 77. Shrestha, Rakesh, et al. ”Anomaly detection based on lstm and autoencoders using federated learning in smart electric grid.” SSRN (2023) 78. Zhang, Weiting, et al. ”Optimizing federated learning in distributed industrial IoT: A multi-agent approach.” IEEE Journal on Selected Areas in Communications 39.12 (2021): 3688-3703. 79. Cao, Keyan, et al. ”An overview on edge computing research.” IEEE access 8 (2020): 85714-85728. 80. Huang, Fang, et al. ”A Federated Domain Adaptation Algorithm Based on Knowledge Distillation and Contrastive Learning.” Wuhan University Journal of Natural Sciences 27.6 (2022b): 499-507. 81. Bokrantz, Jon, et al. ”Maintenance in digitalised manufacturing: Delphi-based scenarios for 2030.” International Journal of Production Economics 191 (2017): 154-169. 82. Ahmadi, Saeed, and Maliheh Salehfar. ”Privacy-preserving cloud computing: ecosystem, life cycle, layered architecture and future roadmap.” arXiv preprint arXiv:2204.11120 (2022). 83. Yang, Fan, et al. ”pDPoSt+ sPBFT: A high performance blockchain-assisted parallel reinforcement learning in industrial edge-cloud collaborative network.” IEEE Transactions on Network and Service Management 20.3 (2022b): 2744-2759. 84. Kumar, Veepan, Prem Vrat, and Ravi Shankar. ”A graph-theoretic approach to evaluate the intensity of barriers in the implementation of Industry 4.0.” International journal of innovation and technology management 18.08 (2021): 2150039. 85. Venkatesan, S., K. Raja Rajeshwari, and M. Ramakrishnan. ”A lightweight searchable encryption and delegation mechanism with forward privacy for improving the security of industrial internet of things-cloud systems.” (2022). 86. Li, Wenjie, et al. ”PBFL: Privacy-Preserving and Byzantine-Robust Federated Learning Empowered Industry 4.0.” IEEE Internet of Things Journal (2023b). 87. Li, Qizhong, Yizheng Yue, and Zhongqi Wang. ”Deep Robust Cramer Shoup delay optimized fully homomorphic for IIOT secured transmission in cloud computing.” Computer Communications 161 (2020): 10-18. 88. Sharma, Durgesh M., Shishir Kumar Shandilya, and Suresh Chandra Satapathy. ”Maximizing blockchain security: Merkle tree hash values generated through advanced vectorized elliptic curve cryptography mechanisms.” Concurrency and Computation: Practice and Experience 35.23 (2023): e7829. 89. Milicic, Ana, et al. ”An autonomous system for PLM domain data exploitation.” International Journal of Computer Integrated Manufacturing 30.1 (2017): 109-120. 90. Campbell, Heather. ”Planning to change the world: Between knowledge and action lies synthesis.” Journal of Planning education and Research 32.2 (2012): 135-146. 91. McMahan, Peter, and Daniel A. McFarland. ”Creative destruction: the structural consequences of scientific curation.” American Sociological Review 86.2 (2021): 341-376. 92. Potts, Colin, and Idris Hsi. ”Abstraction and context in requirements engineering: toward a synthesis.” Annals of Software Engineering 3.1 (1997): 23-61. Information & Authors Information Version history V1 Version 1 30 January 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords healthcare and society privacy security in business Authors Affiliations Bahaa Eltahawy 0000-0001-6372-7547 [email protected] Vaasan yliopisto View all articles by this author Metrics & Citations Metrics Article Usage 421 views 150 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Bahaa Eltahawy. Industrial Systems and Industrial Data Privacy -- A Review. Authorea . 30 January 2025. DOI: https://doi.org/10.22541/au.173822597.78031676/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text {"doi":"10.22541/au.173822597.78031676/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a0245db21c7a1640',t:'MTc3OTg3ODA1NA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();