Certain investigations on improvements in classification algorithms for application layer distributed denial of service attacks

Abstract

Distributed Denial of Service (DDoS) attack is a continuous and newlinecritical threat to the Internet. Traditionally, DDoS attacks are carried out at newlinethe network layer. Recently, there is an increasing number of DDoS attacks newlineagainst online services and Web applications. These attacks are targeting the newlineapplication level. Application Layer-DDoS (AL-DDoS) attacks focus on newlineexhausting the server resources such as Sockets, Central Processing Unit newline(CPU), Memory, Disk/Database bandwidth, and Input/Output (I/O) newlinebandwidth. These attacks are typically more efficient than Transmission newlineControl Protocol (TCP) or User Datagram Protocol (UDP) - based attacks, newlinerequiring fewer network connections to achieve their malicious purposes. It is newlineharder to detect both because it does not involve large amount of traffic and it newlinelooks similar to normal traffic.In this research work, techniques are proposed to generate a newlineSupport Vector Machine (SVM) optimization for predicting AL-DDoS newlineattacks. The proposed work is carried out using three different optimization newlinealgorithms, namely SVM optimized Cuckoo Search (CS), SVM optimized newlineBat algorithm and SVM optimized hybrid CS-Bat algorithm. Investigations newlineare carried out using existing publicly available datasets and a simulated newlinedataset obtained in the cloud running biotechnology application. newline newline