Analisador de alarmes de tráfego de redes através de wavelets

Abstract

In current times, there is a large rise of various services using the internet. And in the same proportion there is an increase of developed methods to achieve benefits of a computer network in an illegal way. Thus, it is necessary to research and develop systems that block improper access. Even setting methods to inhibit the action of malicious users, at any instant they find new vulnerabilities in systems which allow, in some way, compromising the of data security. Security is considered to be every degree of protection from risks relating to internal or external threats that can result in unauthorized access of some information. The key point to the success of a system is to keep the probability low for inaccessible information because of not planned interruptions. Thus, this work is basically about identifing possible attacks to the availability of systems, and afterwards generate results with a low number of false positives. For this, it was necessary to investigate which attacks are covered; what are the main characteristics of each attack; and how this attacks can be detected. To detect an attack, we used an intrusion detection system based on Time Series, however, this IDS presents a high level of false alarms. To minimize the false alarms, we have used a filtering system for the alarms based on wavelets. As result we have get an improvement on IDS reliability.