A Survey on AI-Based Anomaly Detection for Cloud Security in IIoT Environments

Aim: This article examines the cybersecurity weaknesses associated with cloud computing and the relevant legal regulations within the European Union that address these issues. The topic looks into the responsibilities and legal issues surrounding cloud cybersecurity. It explains the roles of cloud actors, such as controllers and processors, and how contracts—such as Service Level Agreements (SLAs)—can help mitigate cybersecurity threats. Finally, the paper addresses the contemporary trends shaping cloud security, such as artificial intelligence (AI) and blockchain technology. Methodology: This article utilizes a doctrinal legal analysis method, systematically reviewing relevant European regulations (GDPR, NIS2 Directive, Cybersecurity Act), contractual frameworks, and academic literature. Sources were selected based on their authority, relevance, and currency, focusing specifically on cybersecurity obligations, liability issues, and emerging technologies like AI and Blockchain. Through comparative analysis and synthesis, the research identifies key legal interpretations and technological impacts on cloud cybersecurity. Findings: Ensuring cybersecurity in the cloud environment is possible, but it remains a complex task. It is a shared responsibility of the cloud parties. However, many gaps and challenges may still exist. That is why employing technological innovations would enhance the security levels in the cloud thanks to the capabilities they offer. Generally, when applied properly, these measures and techniques would improve the overall security of the cloud environment, leading to crucial legal and economic outcomes for the cloud stakeholders in the market. Value: Recently, cloud-computing technology has been evolving as a general-purpose technology whose impact and adoption spans all sectors of the economy. Crucially, its overall cybersecurity is a pivotal concern. While regulations and technologies offer exciting potential pathways for detecting and proving cybersecurity threats and malicious behaviors in the cloud ecosystem, their deployment raises new legal and technical concerns. This article calls attention to the need for several cloud security requirements, such as pre-contractual risk assessments, improved regulatory effectiveness, and the establishment of credible cybersecurity certification systems. Ultimately, this contributes to enhancing the overall security of the cloud environment.

In the rapidly evolving landscape of digital services, cloud computing has become a cornerstone for data storage, processing, and transfer, significantly amplifying security challenges. This environment faces numerous threats and security issues, including cyberattacks, data breaches, and malware, in addition to challenges related to ensuring confidentiality, integrity, and availability. This study aims to highlight the effective role of machine learning techniques in enhancing cloud security by detecting cyber threats, protecting data privacy, and preventing attacks. It underscores the pressing need to adopt intelligent solutions, given the limitations of traditional systems in addressing the complex and evolving threats within cloud environments. Through a critical review of recent research on this topic, the study examines various machine learning methods proven effective in tackling key cloud security issues, such as network intrusion detection, financial fraud detection, and malware identification. The findings confirm that machine learning techniques significantly improve the efficiency of security systems. However, challenges remain, including adapting to continuously evolving threats, ensuring the quality of available training data, and addressing the high computational costs of real-time model operation. The study recommends further research to enhance cloud security, improve adaptability to emerging threats, and establish a theoretical foundation to guide future efforts toward strengthening security in cloud environments.

In this paper, we propose a novel method for the detection of road accidents by analyzing audio streams for road surveillance application. In the last decade, due to the increase in a number of people and transportation vehicles, traffic accidents have become one of a major public issue worldwide. The vast number of injuries and death due to road traffic accident reveals the story of a global crisis of road safety. A total of 52,160 road traffic accidents (RTA), 1130 injuries and 85 fatalities were registered during the year 2000, in the state of Qatar. An increase in the number of transportation vehicles around cities has risen the need for more security and safety in public environments. The most obvious reason for a person's death during accidents is the absence or prolong response time of the first aid facility, which is due to the delay in the information of the accident being reached to the police, hospital or ambulance team. In the last couple of years, several surveillance systems have been proposed based on image and video processing for automatically detecting road accidents and car crashes to ensure a quick response by emergency teams. However, in some situations such as adverse weather conditions or cluttered environments, the visual information is not sufficient enough, whereas analyzing the audio tracks can significantly improve the overall reliability of surveillance systems. In this paper we propose a novel method that automatically identifies hazardous situations such as tire skidding and car crashes in presence of background noises, by analyzing the audio streams. Previous studies show that methods for the detection, estimation, and classification of nonstationary signals can be enhanced by utilizing the time-frequency (TF) characteristics of such signals. The TF-based techniques have been proved to outperform classical techniques based on either time or frequency domains, in analyzing real-life nonstationary signals. Time-frequency distributions (TFDs) give additional information about signals that cannot be extracted from either the time domain or frequency domain representations, i.e. the instantaneous frequency of the components of a signal. In order to utilize this extra information provided by TF domain representation, the proposed approach extracts TF image features from quadratic time-frequency distributions (QTFDs), for the detection of audio anomalies. The extended modified-B distribution (EMBD) is utilized to transform a 1-dimensional audio signal into 2-dimensional TF representation which is interpreted as an image. The image descriptors based features are then extracted from the TF representation to classify visually the audio signals into background or abnormal activity patterns in the TF domain. The proposed features are based on Haralick's texture features extracted from the TF representation of audio signals considered and processed as a textured image. These features are used to characterize and hence classify audio signals into M classes. This research study demonstrates that a TF image pattern recognition approach offers significant advantages over standard signal classification methods that utilize either t-domain only or f-domain only features. The proposed method has been experimentally validated on a large open source database of sounds, including several kinds of background noise. The events to be recognized are superimposed on different background sounds of roads and traffic jam. The obtained results are compared with a recent study, utilizing the same large and complex data set of audio signals, and the same experimental setup. The overall classification results confirm the superior performance of the proposed approach with accuracy improvement up to 6%.

This book covers security of FPGA-accelerated cloud computing environments, with a review of the state-of-the-art in threats as well as defenses.

Security and reliability of cloud computing services remain among the dominant concerns inhibiting their pervasive adaptation. The distributed and the multi-tenancy nature of the cloud computing paradigm can be considered as the root causes for their increased risks and vulnerabilities. Resource sharing and virtualization can also be mentioned as additional main factors contributing to or augmenting cross-site scripting and other cloud vulnerabilities. Cloud are also exposed to the risks and liabilities faced by other networked systems. Poorly designed APIs that may cause security problems or distributed denial of services attacks are the examples of this category that are considered in this paper. Public key infrastructure provides the foundations for provision of some essential security services. These include services such as confidentiality, authentication, and privacy that are of vital importance for establishing trust and confidence between the cloud providers and their clients. In this work, we will discuss the potential flaws of this infrastructure and examine how they may deteriorate the security and reliability levels of the cloud environments. To enable a comprehensive study of the challenges in security and reliability of the cloud computing environments, we categorize the risks and vulnerabilities they face. Traditional techniques, based on cryptography, can address some of these challenges to a certain degree. We will argue that they may not be efficient for use in cloud environments. We then focus on data-centric and homomorphic encryption methods that may provide more appropriate solutions in addressing the challenges in cloud computing security and reliability.

Outsourcing data to a third-party controlled cloud computing services arises various security issues. High security schemes are required to protect data in the cloud platform. Division and Replication of Data in the Cloud for Optimal Performance and Security (DROPS) method addressed these issues by improving the security and performance of a cloud environment. An enhanced version of DROPS named SePeCloud was proposed that improved the security and performance hand-in-hand with Fog based Deduplication and Privacy Preserving Online Updating. This paper improves the security and data storage with Self-Destruction mechanism by handling the applications independently built on reinforcement learning strategy. Data chunk fingerprint index and sketch index are included to support independent and parallel data destruction among multiple applications. Additionally, SePeCloud is further extended to prevent impersonation attacks from illegitimate users by introducing Modified Shamir secrete sharing scheme to handle user revocation policies with limited storage space. The experimental results prove that the final version of the SePeCloud performed better than the previous versions in terms of Replication cost savings and Computation time by improving both security and performance of the cloud system.

In this chapter, we are going to take a top-down approach and think how to properly organize your cloud resources around projects, environments, and solutions you are about to design and run on Oracle Cloud Infrastructure.

While paperless is a source of tremendous opportunities for companies, it is also a bearing of many new risks. Indeed, externalization of electronic filing system can expose the company to several vulnerabilities and threats. We propose, in our gSafe (Government Safe) project a new Cloud service, named Safe Box as a service (SBaaS). SBaaS is used for probative value archiving sensitive documents during a defined period in a secure environment. We propose a layered model that intends to satisfy Cloud user requirements and Cloud security challenges. In this paper we define the technical architecture of the service and its basic components. In addition we discuss its feasibility using Hadoop Distributed File System.

ABSTRACTAnomaly detection is an active research topic in hyperspectral remote sensing and has been applied in many areas including environmental monitoring, urban survey, mineral mapping, and national security. Usually, it makes a detection decision without any prior target information or background information. Several anomaly detection algorithms (e.g. Reed–Xiaoli detector, blocked adaptive computationally efficient outlier nominator and random-selection-based anomaly detector) have been developed, which rely on estimating background information only from a hyperspectral image without considering target information in making a detection decision. These methods may be efficient in general but sometimes with high false alarm rate (FAR). In order to reduce FAR, this study proposes a novel method that incorporates both background and target information, derived from the hyperspectral imagery, into anomaly detection algorithms. The target information is helpful to detect anomalies as outliers. With a scene of real airborne visible infrared imaging spectrometer data, the experimental results demonstrate that the proposed method has produced better detection results and higher time efficiency compared with those using the traditional algorithms that only consider background information.

The financial benefit of a refined, intelligent power system contain rising demand is known as a smart grid (SG). Energy use has increased significantly due to new technology and population growth, creating significant issues for the environment and energy security. Utilizing the Deep Learning (DL) Method in conjunction with power management is important for issue solving. Data is pre-processed using a normalization method after being gathered via a smart city in a power-hungry smart grid. The research used a Discrete Fourier Transform (DFT) to extract features and estimate smart city power usage within a smart grid setting utilizing large-scale, high-dimensional data. The research introduced a unique DL-based, Hybrid Mountain Gazelle Optimized Graph Neural Networks (HMGO-GNN) for real-time forecast and adaptive management of smart grid stability. The suggested technique can stabilize critical machines; predict transient stability, and the stability limit. It can also be used to detect instability during the first swing and multiple times. Experiments conducted on simulation environments illustrate that the intended method has a high precision to predict system stability in real time. Evaluation of performance with the help of essential metrics of evaluation, i.e., F1-score (96%), precision (94%), accuracy (97%), AUC (98%), and recall (95%), presents a holistic understanding of whether or not the model succeeds in forecasting grid instability instances. The cost benefits, decreased energy consumption, and remarkable efforts in decreasing greenhouse gas emissions eventually contribute towards increased environmental sustainability. Additionally, the adaptability and flexibility of the framework in implementation for use in other urban situations are also exemplified. The study accentuates the ability of the framework to enhance efficiency, facilitate sustainability, and provide sophisticated solutions for the current challenges in energy across the globe.

Rare diseases affect approximately 1 in 11 Americans, yet their diagnosis remains challenging due to limited clinical evidence, low awareness, and lack of definitive treatments. Our project aims to accelerate rare disease diagnosis by developing a comprehensive informatics framework leveraging data mining, semantic web technologies, deep learning, and graph-based embedding techniques. However, our on-premises computational infrastructure faces significant challenges in scalability, maintenance, and collaboration. This study focuses on developing and evaluating a cloud-based computing infrastructure to address these challenges. By migrating to a scalable, secure, and collaborative cloud environment, we aim to enhance data integration, support advanced predictive modeling for differential diagnoses, and facilitate widespread dissemination of research findings to stakeholders, the research community, and the public and also proposed a facilitated through a reliable, standardized workflow designed to ensure minimal disruption and maintain data integrity for existing research project.

With the rapid development of cloud computing technology, authentication and authorization management have become core elements for ensuring cloud environment security. While existing surveys have covered various aspects of cloud security, few have integrated region-specific compliance requirements with a quantitative analysis of user experience. This paper provides an in-depth review of authentication mechanisms and fine-grained authorization strategies in cloud computing, with a particular focus on policy-based methods. The main contributions of this paper are threefold: First, it presents a novel synthesis that incorporates the compliance requirements of Chinese Classified Protection and Commercial Cryptography into the cloud authentication and authorization framework, offering a valuable perspective for practitioners in regulated industries. Second, it introduces a set of quantitative metrics to evaluate the usability and effectiveness of various authentication mechanisms, moving beyond purely qualitative descriptions. Third, it provides a detailed comparative analysis of access control models (RBAC, ABAC, PBAC) and discusses their applicability in modern, dynamic cloud environments. By elaborating on technological principles, practical applications, and future trends, this paper aims to provide comprehensive reference and guidance for researchers, developers, and users, thereby helping to build more secure and reliable cloud environments.

Cloud computing and containerized environments have become foundational components of modern IT infrastructure, offering scalability and agility. However, their dynamic nature introduces significant security challenges, including anomalies in traffic, DDoS attacks, hidden crypto mining (cryptojacking), and credential compromise. Traditional signature-based security mechanisms often fail to address these rapidly evolving threats effectively. The objective of this study is to assess the potential of artificial intelligence (AI) and machine learning (ML) in enhancing cloud and container security. Specifically, it explores the effectiveness of AI/ML models for anomaly detection, threat classification, cryptojacking, and DDoS identification, deception-based defenses, and false positive reduction. The methodology involves a structured literature review of key scientific publications from 2023 to 2025. Comparative analysis is conducted on experimental solutions, including hybrid models (XGBoost, CNN, LSTM) in intrusion detection systems; eBPF-based syscalls tracing for container behavior profiling; ML classifiers for vulnerability prioritization in DevSecOps; and active defense platforms combining honeypots and adaptive monitoring loops (MAPE-K). Findings indicate that AI-powered security systems achieve detection accuracies above 99%, reduce false positive rates to around 2%, and enable real-time responsiveness without degrading system performance. Notably, systems that integrate multiple models and utilize low-level data (e.g., syscalls, network patterns) exhibit superior threat identification and resilience. In conclusion, integrating AI into cloud security architectures is essential for ensuring continuous and proactive defense in dynamic infrastructures. The paper also outlines future research challenges, such as the need for explainable AI (XAI), limited availability of high-quality training datasets, and vulnerability to adversarial inputs. The insights are relevant for cybersecurity researchers and practitioners seeking to deploy intelligent defense mechanisms in cloud-native ecosystems.

The rapid adoption of cloud computing has witnessed unprecedented enhancements in data storage capacities, scalability, and operation flexibility for diverse organizations. Yet, the transition from legacy on-premises to cloud-based environments has raised grave concerns about information safety and compliance with regulations in evolving regulatory environments. This research will discuss the most significant challenges and innovative solutions for the deployment of data security mechanisms in cloud environments, with special reference to regulatory compliance from 2015 to 2024. Though cloud vendors have taken major leaps in enhancing the security of cloud centers, some of the longstanding issues like threats to multi-tenancy, poor encryption methods, vendor lock-in situations, and jurisdictional issues continue to create significant problems. In addition, with the regulatory environment in flux, organizations are also facing challenges in terms of compliance with regulations like GDPR, HIPAA, and CCPA in the context of cloud computing. This research tries to determine the gaps in existing literature regarding the integration of sophisticated technologies, including artificial intelligence (AI), blockchain, and machine learning (ML), into cloud security frameworks. In addition, it addresses the organizational challenges in ensuring continuous compliance and adequate data protection, especially in hybrid and multi-cloud settings. The research further explores the limitations of conventional security controls, emphasizing the requirement for automated compliance solutions, strong encryption techniques, and new security models such as Zero Trust and federated identity management.

Security in cloud environments is paramount due to the increasing prevalence of Distributed Denial of Service (DDoS) attacks, which can severely disrupt services and cause significant financial and data losses. Traditional security mechanisms, such as Intrusion Detection Systems (IDS) and Firewalls, often struggle to detect and mitigate novel, evolving threats due to the lack of predefined detection signatures. This paper proposes a real-time signature-based detection mechanism tailored specifically for cloud environments. The proposed system generates signatures in real-time, enabling the identification and prevention of emerging DDoS attacks. The effectiveness of the solution is validated through extensive experimental evaluations, demonstrating its ability to reduce attack impact and enhance cloud security.