idMAS-SQL: Intrusion Detection Based on MAS to Detect and Block SQL injection through data mining

  • Authors:
  • Cristian I. PinzóN;Juan F. De Paz;ÁLvaro Herrero;Emilio Corchado;Javier Bajo;Juan M. Corchado

  • Affiliations:
  • Faculty of Computer Systems Engineering, Technological University of Panama, Building No. 3, Campus Dr. Víctor Levi Sasso, Universidad Tecnológica Ave., Panamá City, Panama;Departamento de Informática y Automática, Universidad de Salamanca, Plaza de la Merced s/n, 37008 Salamanca, Spain;Department of Civil Engineering, University of Burgos, C/Francisco de Vitoria s/n, 09006 Burgos, Spain;Departamento de Informática y Automática, Universidad de Salamanca, Plaza de la Merced s/n, 37008 Salamanca, Spain;Departamento de Informática y Automática, Universidad de Salamanca, Plaza de la Merced s/n, 37008 Salamanca, Spain;Departamento de Informática y Automática, Universidad de Salamanca, Plaza de la Merced s/n, 37008 Salamanca, Spain

  • Venue:
  • Information Sciences: an International Journal
  • Year:
  • 2013

Quantified Score

Hi-index 0.07

Visualization

Abstract

This study presents a multiagent architecture aimed at detecting SQL injection attacks, which are one of the most prevalent threats for modern databases. The proposed architecture is based on a hierarchical and distributed strategy where the functionalities are structured on layers. SQL-injection attacks, one of the most dangerous attacks to online databases, are the focus of this research. The agents in each one of the layers are specialized in specific tasks, such as data gathering, data classification, and visualization. This study presents two key agents under a hybrid architecture: a classifier agent that incorporates a Case-Based Reasoning engine employing advanced algorithms in the reasoning cycle stages, and a visualizer agent that integrates several techniques to facilitate the visual analysis of suspicious queries. The former incorporates a new classification model based on a mixture of a neural network and a Support Vector Machine in order to classify SQL queries in a reliable way. The latter combines clustering and neural projection techniques to support the visual analysis and identification of target attacks. The proposed approach was tested in a real-traffic case study and its experimental results, which validate the performance of the proposed approach, are presented in this paper.