ADVANCED RANDOM TIME QUEUE BLOCKING WITH TRAFFIC PREDICTION FOR DEFENSE OF LOW-RATE DOS ATTACKS AGAINST APPLICATION SERVERS

Abstract

Among many strategies of Denial of Services, low-rate traffic denial-of-service (DoS) attacks are more significant. This strategy denies the services of a network by detection of the vulnerabilities in performance of the application. In this research, an efficient defence methodology is developed against low-rate DoS attack in the application servers. Though, the Improved Random Time Queue Blocking (IRTQB) technique can eliminate the vulnerabilities in the network and also avoiding the attacker from capturing all the server queue positions by defining a spatial similarity metric (SSM). However, the differentiation of the attack requests from the legitimate users’ is not always efficient since only the source IP addresses and the record timestamp are considered in the SSM. It was improved by using Advanced Random Time Queue Blocking (ARTQB) scheme that employed Bandwidth utilization of attacker in IRTQB to detect the DoS attack that normally consumes a huge number of resources of the server. However, this method becomes ineffective when the attack consumes more network traffic. In this paper, an efficient detection technique called Advanced Random Time Queue Blocking with Traffic Prediction (ARTQB-TP) is proposed for defining SSM which contains, Source IP, timestamp, Bandwidth between two requests and the difference between the attack traffic and legitimate traffic. The ARTQB-TP technique is utilized to reduce the attack efficiency in 18 different server configurations which are more vulnerable to the DoS attacks and where the attacks may also have a chance to improve its effectiveness. Experimental results show that the proposed system performs better protection of application servers against the LRDoS attacks by solving its impacts on any kind of server architectures and reduced the attack efficiencies of all the types of attack strategies.