International Journal of Scientific Research in Network Security and Communication
Not a member yet
272 research outputs found
Sort by
Security for AI and IoT Convergence: Novel Perspectives
Get PDFThe conjunction of Artificial Intelligence (AI) and the Internet of Things (IoT) presents a transformative synergy that holds immense promise for various domains, ranging from healthcare and smart cities to industrial automation and autonomous vehicles. However, this convergence also introduces a plethora of security challenges that demand innovative and novel perspectives to safeguard the integrity, confidentiality, and availability of data and systems. This paper explores the intricate landscape of "Security for AI and IoT Convergence" and introduces pioneering approaches and insights to mitigate the evolving threat landscape. Through a comprehensive literature review, we identify the current security challenges inherent in the intersection of AI and IoT, including vulnerabilities in connected devices, data privacy concerns, and the complex interplay between autonomous decision-making and real-time threat detection. We then present novel perspectives and methodologies that leverage cutting-edge technologies like machine learning, Blockchain, and interdisciplinary collaborations to address these challenges effectively. To ground our discussions, we offer real-world case studies that illustrate the practical implementation and impact of these novel security perspectives. We also delve into the evaluation metrics and considerations required to assess the efficacy of these security solutions. Additionally, we highlight the significance of on-going research, regulatory compliance, and ethical dimensions in shaping the future of AI and IoT convergence security. This paper not only serves as an essential reference for researchers and practitioners in the field but also underscores the imperative nature of continuous innovation and vigilance in ensuring the secure coexistence of AI and IoT technologies.
 
Three Hash Functions Comparison on Digital Holy Quran Integrity Verification
Get PDFThis paper provides a study of hash functions comparison on holy Quran integrity verification to ensure integrity of data in digital copy of holy Quran. There are many methods available in data security area for integrity verification. This study presents a comparison of three cryptographic hash functions SHA256, RIPEMD160 and Blake3 which used to verify the integrity of digital holy Quran and determine which one is better. Blake3 hash function is chosen in this proposed schema because it has many characteristics, the most important of which is the speed characteristic, and it will be mentioned in detail in Background section. Furthermore security analysis and performance analysis are focused on this study. In security analysis by apply four experiments to find out the strength and effectiveness of each three hash functions and all possible possibilities of hash collisions be carefully analysed and studied. Speed in term of execution time is measuring for all three hash function and this is under the umbrella of the performance analysis. The results that we have obtained say that the Blake3 hash function is the fastest function with the rest two hash functions. And the resulting number of the time that the attacker need to come up with a fake verse using SHA256/BLAKE3 will be greater than the number obtained by using RIPEMD160.
 
A Review of the Security Architecture for SDN in Light of Its Security Issues
Get PDFSoftware-defined networking (SDN) simplifies and enhances network management for administrators. It is a modern approach that liberates networks from conventional constraints and is considered a game-changing concept for the future of the web. SDN can detect and prevent malicious traffic through its three-tiered architecture, consisting of a control layer , application layer and a forwarding layer. Our work distinguishes between DoS and DDoS attacks, which pose varying degrees of security risk for SDN users. SDN-based 5G networks are vulnerable to DDoS assaults by malevolent users, but can be secured using tools like Mininet. The SDN security architecture can be improved by methods such as network monitoring, verification, automation, improvised threat detection, and dynamic reaction. By separating network control and data planes and using software applications, SDN can effectively detect and stop malicious traffic. Combining SDN and ML enables the detection and prevention of low-rate DoS attacks, providing a security solution for SDN-based 5G networks. The coming model enhances efficacy in detecting and protecting against DoS and DDoS attacks, allowing enterprises to defend their networks and crucial services.
 
Unlocking the Potential of Open and Distance Learning with Artificial Intelligence
Get PDFThe integration of Artificial Intelligence (AI) in the field of education presents a vast array of possibilities, particularly for Open and Distance Learning (ODL) institutions. AI offers various ways for open universities to tackle challenges in areas such as learning, teaching effectiveness, and the advantages and limitations of computer-based systems in education, which are crucial for ODL institutions that heavily rely on human-machine interactions. Indira Gandhi Open University in India, has been utilizing information and communication technologies (ICTs), including AI, to enhance its operations and services for the past seven years. Although AI has not yet been fully implemented in education, IGNOU recognizes its potential to improve quality, pedagogy, and the overall teaching and learning experience. The purpose of this study is to explore how AI can be utilized in ODL institutions, with a focus on expert systems for program advising, automated class scheduling, assignment grading, plagiarism detection, learner retention and adaptation to diverse needs and backgrounds, property maintenance, and security. IGNOU is optimistic that AI has the potential to bring about a significant and positive transformation in ODL and shape the future of open and distance learners.
 
Unlocking Network Security and QoS: The Fusion of SDN, IoT, and Machine Learning: A Comprehensive Analysis
Get PDFThe convergence of Software-Defined Networking (SDN) and the Internet of Things (IoT) has ushered in transformative changes, offering unparalleled levels of network flexibility, programmability, and connectivity. While this integration provides numerous benefits, it also introduces security challenges. Motivated by the imperative to fortify the security posture in this dynamically evolving landscape, this review paper explores the vulnerabilities, threats, and corresponding responses in the security landscape of SDN and IoT. Recognizing the critical need for proactive security measures, the paper underscores the potential of Quality of Service (QoS) empowered by Machine Learning (ML) as a solution. By harnessing ML, QoS emerges as a powerful means to proactively identify and mitigate potential attacks, offering an effective approach to enhance network security. The motivation behind integrating QoS with ML lies in its ability to ensure dependability, availability, and integrity, thereby instilling confidence in the reliability and resilience of the interconnected world. The paper goes through examination of challenges, delving into the proactive management of QoS within SDN, intricacies of IoT network architectures, and the unique features and limitations of IoT systems. Furthermore, it comprehensively addresses potential countermeasures for various security threats, such as Denial of Service (DOS), Man-in-the-Middle (MITM) attacks, and Ransomware attacks, particularly on devices with limited resources. This abstract provides a concise yet comprehensive overview of the paper`s motivations, emphasizing the urgency and significance of the proposed solutions for securing modern network environments
Credit Risk Assessment in Non-Banking Financial Institutions: Lessons from Shadow Banking Sector
Get PDFThis research paper explores the credit risk assessment practices in non-banking financial institutions (NBFIs) with a focus on lessons learned from the shadow banking sector. NBFIs have gained significant prominence in the financial landscape, and their role in credit intermediation has expanded. However, the inherent complexities and unique characteristics of NBFIs pose challenges to credit risk assessment. Drawing insights from the shadow banking sector, this study aims to identify key lessons and best practices that can enhance credit risk assessment in NBFIs. The research adopts a qualitative approach, analyzing relevant literature, regulatory frameworks, and case studies to develop a comprehensive understanding of credit risk assessment practices in NBFIs. The findings highlight the importance of robust risk management frameworks, adequate risk governance, effective monitoring mechanisms, and the use of innovative tools and technologies in mitigating credit risks in NBFIs. The research concludes by providing recommendations for policymakers, regulators, and NBFIs to strengthen credit risk assessment practices and ensure the stability and resilience of the financial system.
 
A Review of Credit Card Fraud Detection Using Machine Learning
Get PDFNowadays fraud has been increasing due to the establishment of online payment mode on different E-commerce platform.A credit card is a form of payment that lets you buy goods or services on credit from an issuer, usually a bank. You can make purchases up to a specified limit and then pay them off over time either in full or with minimum payments.There are several types of security features including fraud protection, verified by visa and master card secure code, address verification systems, and biometric authentication. Additionally, some cards offer the additional security feature of a chip and pin system which requires that the cardholder enter a secret code to make purchases.Still fraud has been executed using this card. In this fraud, banks, merchants, and organisations are losing billions of dollars. According to one survey, the prevalence of credit card fraud is rising by 12.5% a year. It is crucial to identify fraud using secure and effective methods. Nowadays, hybrid algorithms and artificial neural networks are used to detect fraud since they perform better than other methods. We will use dataset variables like "duration," "amount of transaction," and "V1 to V28" as derived parameters for this. We will build a model that will separate out fraudulent transactions from other transactions using machine learning techniques or algorithms.
 
A Review on New Multilevel Scheduling Algorithm and SJF and Priority Scheduling Algorithms
Get PDFThe two CPU scheduling algorithms that received the majority of our attention in this paper after reviewing a variety of CPU scheduling algorithms were the shortest job first and priority scheduling algorithms, as well as an improved priority scheduling algorithm that performs better than current scheduling algorithms. Scheduling is the process of assigning tasks to the CPU to optimize use. Because the CPU is the most important resource in a computer system, numerous scheduling approaches aim to maximize its use. The purpose of this study is to explore the CPU scheduler`s construction of high-quality scheduling algorithms that meet the scheduling objectives and to study the performance of a multilevel scheduling algorithm that combines two scheduling algorithms.
 
Implementation of Intrusion Detection and Prevention System Based on Software Approachs
Get PDFThe detection of security vulnerabilities is ever more difficult as cyber-attacks get more complicated. The faith in security services like data confidentiality and integrity is eroded by a failure to prevent security breaches. The literature has suggested a number of intrusion detection techniques to counter computer security risks. This effort aims to create an intrusion detection and prevention system, or "IDPS,” An integrated system that maximizes each factor`s advantages while reducing its disadvantages is the proposed solution for intrusion prevention. By demonstrating how attackers can elude detection. The outcome of this research is a security system that can recognize attack attempts, block the IP address of the attacker, and carry out network forensic investigations. According to the findings of our study, Snort the IPS mode in PfSense, can identify assaults aimed at your system, and PfSense, having visualization capability, immediately implements preventive actions by blocking the attacker`s IP address. Network forensics can use this method to conduct an investigation into an attack and determine whether the attack is having a negative impact based on the alarms produced by the snort. It also sheds light on potential future research challenges to stop these attacks and strengthen computer systems` security.
 
Design and Analysis of Hybrid Software Defined Network System for Stability Enhancement
Get PDFSoftware-defined networks (SDNs) facilitate more efficient routing when traffic flows using centralized network view. On the other hand, traditional distributed routing still has the advantage of better scalability, robustness, and swift reaction to events such as failure. Therefore significant potential benefits to adopt a hybrid operation where both distributed and centralized routing mechanisms co-exist. This hybrid operation however imposes a new challenge to network stability since a poor and inconsistent design can lead to repeated route switching when the two control mechanisms take turns to adjust the routes. In this paper, we discuss the ways of solving the stability problem. To develop the stable networking environment three tire routing architecture is proposed. The router stability is enhanced and speed of data transfer is high in this proposed system. The router will analyze the network traffic that occurs during the data transfer and redirect to the other router. The algorithms used in this proposed systems are Support vector machine algorithm and Multilayer perception algorithm. The stability of a hybrid-software defined network involves the consistency between the centralized routing performed by the centralized controller and the distributed routing performed by the individual local routers. If these two control units are not consistent with each other, the routing decision may be overturned repeatedly as they take turns to modify the routes.