Efficient denial of service attack resistance and prevention using adaptive caching system in hybrid network

Abstract

Distributed Denial of Service (DDoS) attacks are the most commonly occurred attacks in a network which affects the services provided by large application servers. Such DDoS attacks are formed by making legitimate requests to frequently visited websites. Hence, these DDoS attacks are considered as an important security issue which is to be monitored in a network for preventing flooding attack. By preventing these attacks, confidence level of a network is improved efficiently with the improvement ofresponse rate, true positive rate, and throughput and attack resistance rate. Communication is established to the users with the help of internet by performing information exchange using social media. At the time of communication, providing security level and authentication of users is a challenging task to prevent from attackers who make malicious traffic with fake identity. Denial of service (DoS) attack is a type of explicit attack made by attacker to acquire resources of authenticated users. Some of DoS attacks include flash crowd, flooding, teardrop, etc. Flash crowd in internet is a type of DDoS attack that affects services provided by server for users. This flash crowd attacks are resulted by legitimate form of requests from the side of users of often visited websites. Traditional process of identifying Distributed Denial of Service (DDoS) attack in internet employed entropy variations that presented a trace back method between usual and DDoS traffic. But trace back method wrongly treated a flash crowd as a DDoS attack and hence produces false positive alarms. Also, discrimination between DDoS attack and flash crowd is not addressed effectively. An embedded Markov Chain is presented to examine the rule-based network firewalls. Extraction of key features is carried out to distinguish normal traffic and DDoS attacks that targeting different rule positions. But the mitigation solution for affected firewalls in real time is not considered. Vulnerability of Network Mechanisms is analyzed for monitoring the performance degradation of a system which is caused by malicious users whose aim is consuming abundance resources. However, Hash tablebased Vulnerability of Network Mechanisms does not solve the problem of longer waiting time based on attack size. On the other hand, a Queuing model based on Embedded Markov Chain is employed for analyzing rule-based firewalls which have standard network traffic flows as well as DoS attack flows. This queuing model targets several rule positions but mitigation of DoS attacks targeting bottom rules are achieved only with minimal confidence in the network. newline