Application of PSVR-DNS Algorithm for Attacker Detection and Isolation

Abstract

The DNS (Domain Name System) is used to map and convert human-friendly domain names to the numeric IP (Internet Protocol) addresses. As with the operation of any communication system, there are some security risks associated with the operation of DNS. Actions targeting the availability or stability of a network's DNS service are considered DNS attack. For example, a high volume of traffic and a large number of requests coming to DNS servers are part of a type of DoS (Denial of Service) attack that uses DNS for amplification. Although most DNS servers are open source, some commercial protective DNS services are available for network traffic control, filtering and automatic blocking of requests to undesirable, dangerous or malicious internet domains, but the price of such services is high. In this paper, a new PSVR-DNS (Probability Support Vector Regression-Domain Name System) algorithm is proposed for the purpose of detecting and isolating attackers who pose a threat to an uninterrupted work of the DNS servers. The main focus is on the prevention of the DNS cache poisoning. The collected results showed that the proposed PSVR-DNS algorithm achieves better performance related to faster detection and isolation of attacks compared to some existing algorithms.

Keywords

• Bayesian probability
• DoS attack
• PSVR-DNS algorithm
• SDN