K. Munivara Prasad ¹, A. Rama Mohan Reddy ², K. Venugopal Rao ³

Affiliations
1 Department of CSE, JNTUH, Hyderabad - 500085, Telangana, IN
2 Department of CSE, SVUCE, SV University, Tirupati - 517502, Andhra Pradesh, IN
3 Department of CSE, GNITS, Hyderabad - 500008, Telangana, IN

Abstract

To devise an Anomaly based Real Time Prevention (ARTP) of under rated App-DDOS attacks on Web for achieving fast and early detection. Method: We proposed a model based on machine learning approach that used to achieve the fast and early detection of the App-DDOS by multitude request flood. The proposed model ARTP is focused on defining set of metrics called "Re-quest chain length, request chain context, ratio of packet types, ratio of packet count, route context, router chain context and ratio of request intervals. The key factor of the proposal is unlike many of the bench marking models, which are considering requests or sessions as input to discover the anomalies, it considers set of requests are sessions in a time frame discovered to identify the anomalies of the metrics proposed. The experiments were carried out on bench marking LLDOS dataset and the performance analysis was done by the statistical analysis of the metrics like precision, recall, sensitivity and specificity. The process over-head also assessed in order to estimate the scalability and robustness of the proposal. Findings: The proposed model is highly significant in App-DDOS attack detection to adopt by current scenario of web applications with crowded requests that is phenomenally magnified to petabytes that compared to the past web request load in gigabytes.

Keywords

APP-DDoS, ARTP, Distributed Denial of Service, DDoS Atacks, HTTP Flooding, Intrusion Detection.

Full Text