Adaptive Detection of Network Dos and DDOS Attacks Using Sampling and Machine Learning
- Version
- Download 1
- File Size 444.82 KB
- File Count 1
- Create Date 6 June 2025
- Last Updated 6 June 2025
Adaptive Detection of Network Dos and DDOS Attacks Using Sampling and Machine Learning
S. Sasikumar1, K Nivetha2
1 Student, Master of Computer Application, Dr.M.G.R. Educational and Research Institute, Chennai, Tamil Nadu, India.
2 Professor, Master of Computer Application, Dr.M.G.R. Educational and Research Institute, Chennai, Tamil Nadu, India.
ABSTRACT
Distributed Denial-of-Service (DDoS) attacks have become a major concern in modern network security, threatening the availability and integrity of services across various platforms. These attacks flood the target network with a high volume of traffic, often overwhelming the system and causing it to crash or become unresponsive. As the scale and complexity of DDoS attacks continue to grow, traditional detection methods, such as signature-based systems or threshold-based approaches, are proving inadequate. These methods frequently suffer from high false positives and delays in identifying attacks, which can result in significant damage or downtime before corrective measures are implemented.To address these challenges, this project proposes the development of an Adaptive Detection System (ADS) for detecting and mitigating network DoS and DDoS attacks. The system leverages advanced sampling techniques and machine learning (ML) algorithms to dynamically analyze network traffic and identify malicious patterns more accurately. Unlike traditional methods, the proposed system can adapt to the ever-evolving nature of cyberattacks, reducing the likelihood of missed detections or false alarms. This estimation helps in preventing service disruptions while ensuring the system remains robust. Additionally, the project aims to reduce computational complexity by utilizing the properties of the Kronecker product, allowing for more efficient processing of large-scale network data.An observer-based model is also introduced, which enhances the overall system’s resilience against attack impacts. This model establishes a consensus criterion, ensuring that even in the presence of attackers, the system can maintain
Download