International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
ENHANCING DDOS DETECTION IN IOT SYSTEMS THROUGH BOOSTING TECHNIQUES POORNIMA@ PRIYANKA.R1, KARUPPURAJA.S2, MOHAN RAJ.P3,LAVANESH.T.M4 1Assisstant Professor 1, Department of Computer Science and Engineering,
K.L.N. College of Engineering, Sivagangai, India
2,3,4,Student, Department of Computer Science and Engineering,
K.L.N. College of Engineering, Sivagangai, India
---------------------------------------------------------------------------***-------------------------------------------------------------------------
Abstract-Distributed Denial of Service (DDoS) attacks
develop datasets and machine learning models for effective attack detection and feature selection.
pose in both traditional networks and in Internet of Things (IoT), including home networks. This paper proposes the application of DDoS traffic detection model leveraging boosting algorithms to address this critical issue. With the IoT devices, attackers exploit vulnerabilities of botnet networks, amplifying the impact of DDoS attacks. Machine learning techniques algorithms is used for DDoS detection, categorized into supervised and unsupervised approaches. Despite advancements in Machine Learning (ML) and deep learning, DDoS attacks remain a significant threat to the integrity and availability of the Internet. Our proposed model utilizes boosting learning classification algorithms to analyze network data and identify malicious traffic patterns. Evaluation of the detection model relies on publicly available datasets, ensuring robustness and generalizability. The primary objective of this project is to develop effective algorithms for identifying and mitigating DDoS attacks within networks. Additionally, as social networks continue to grow exponentially, detecting attacks within these platforms presents a complex challenge. To address this, our research involves the development and comparison of four distinct machine and deep learning algorithms for DDoS detection. Through this research, we aim to contribute to the ongoing efforts to safeguard network against DDoS attacks, thereby enhancing the security of the Internet ecosystem.
Distributed Denial of Servicethreats involves deploying machine learning-based detection frameworks, particularly for botnet attacks prevalent in IoT environments. The Botnets, controlled by a single entity, can consist of thousands of compromised devices operating covertly. Machine Learning techniques and Deep Learning, offer solutions for network detection in IoT systems. However, the IoT devices require efficient feature selection and lightweight detection systems to detect high-dimensional traffic data and sophisticated attacks. UNWS dataset aim to provide labeled data for supervised learning, essential for accurate detection in IoT networks. The previous Researchers developed a framework for predicting DDoS attack types using machine learning, employing Random Forest and XG-Boost algorithms. They utilized the UNWS-np-15 dataset, obtained from GitHub, and implemented the framework in Python. After model deployment, performance is evaluated using confusion matrices. Results indicated Random Forest achieved Precision and Recall ofapproximately 89%, with an average Accuracy of around 89%. XG-Boost yielded Precision and Recall of about 90%, with an average Accuracy of approximately 90%. Comparatively, their model significantly outperformed existing research, which reported defect determination accuracies of approximately 85% and 79%.
KeyWords-Distributed Denial of Service attacks, Internet of Things, Boosting algorithms, Machine learning techniques, Four distinct machine and deep learning algorithms
This system utilizes the UNWS dataset sourced from its repository as input. Data preprocessing is conducted to handle missing values and encode input data labels, essential for accurate predictions. The dataset is then divided into training and testing sets based on a specified ratio, with the majority allocated to training and a smaller portion to testing. Training data is used to evaluate the model, while testing data is employed for prediction. Classification algorithms, including both machine learning
1.INTRODUCTION IoT devices are increasingly targeted by attackers due to vulnerabilities in their software and hardware, leading to the creation of botnets for DDoS attacks. As IoT devices in homes and organizations, their minimal security measures make them easy targets. Stakeholders, including users and service providers, have a vested interest in preventing IoT devices from being exploited. Efforts are underway to
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 2618