A Novel Framework for Detection of DoS/DDoS Attack Using Deep Learning Techniques, and An Approach to Mitigate the Impact of DoS/DDoS attack in Network Environment
Year: 2023
Type: article
Abstract: DoS attacks are a major network security issue. The internet and computer networks are essential to our daily lives and businesses. With our reliance on computers and communication networks, harmful actions have increased. Network hazards plague modern communication. To keep networks running smoothly and users' data safe, network traffic flow must be monitored for malicious activity and assaults. Denial-of-service (DoS) attacks aim to disrupt a network server, website, or web service. Computer networks and services are vulnerable to DoS and DDoS attacks. Flooding may be the simplest DDoS assault. DDoS attacks transmit massive amounts of useless data to a network or server. The study seeks to strengthen network infrastructures against various threats, maintain service continuity, and secure the network. Denial-of-service (DoS) attacks prevent legitimate users from accessing and using information systems and resources. Figure B shows DoS/DDoS attacks using ICMP, UDP, and the more prevalent TCP flood assaults. These strikes must be detected and stopped immediately. Businesses and schools went online during COVID-19. Because so much data is created and stored, traditional Machine Learning-based DoS/DDoS attack detection approaches are ineffective. This study uses SVM, MLP, and LSTM algorithms for Deep Learning. The proposed Deep Learning model learns and builds binary and multiclass classification models that can distinguish network attack activity from normal traffic. We look for outliers and attack signals in traffic patterns and data. Our deep learning model is studied with accuracy and precision. In detection, the system checks for attack or regular network data. MLP Algorithm helps this model discover items 97% of the time. LSVM ML classification compares the suggested system's performance. This paper examines traffic behavior. This study also used traffic filtering to eliminate suspicious or attack-signature traffic. Next, we limited traffic from specified sources and locations using rate-limiting. Python SCAPY and wireshark Sniffer in Linux OS capture network packet data for analysis and repair. Compared wireshark with scapy packet capturing analysis and mitigation. This study examines network DoS/DDoS assaults and their prevention. These approaches detect and mitigate flood-based DoS assaults to keep systems functioning and networks safe. To keep up with DoS assaults and the threat landscape, you must continually studying and developing new tactics.
Institution: Jawaharlal Nehru Technological University, Kakinada
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7218068088
