BIFAD: Bio-Inspired Anomaly Based HTTP-Flood Attack Detection

Application layer based DDoS attacks have changed the way DoS attacks are taking place with more subtle level of attacking methods being imparted, which pose an ever-increasing challenge towards the emerging trends of internet based application systems development. Among the key range of attacks that take place, HTTP flood DDoS attacks are on high. In the case of DDoS attacks based on HTTP flood, unusual quantum of requests are sent to the servers within quick time interval and it affects the response and the performance levels of the server . There are numerous solutions in contemporary literature, pertaining to thwarting HTTP flood kind of attacks. It is imperative from the analysis that there are constraints in the existing models since the most of these models are user session based and/or packet flow patterns. The session based evolution models are vulnerable to botnets and packet flow pattern based models are vulnerable if attack sources are equipped with human resource and/or proxy servers. Hence, there is inherent need for improving the solutions towards addressing the HTTP flood kind of attacks over the system. The crux for such system is about ensuring that fast and early detection with minimal false alarming in streaming network transactions, and ensures that the genuine requests are not impacted. To address such a system, the model of Bio-Inspired Anomaly based HTTP-flood detection aimed, and the proposed model depicted in detail along with experimental inputs. Results attained from the process exemplify the significance and robustness of the model towards achieving the objectives considered for the solution.