Publication: Sistema de detección de ataques DDoS en Tor
Loading...
Official URL
Full text at PDC
Publication Date
2015
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
El presente trabajo trata de hacer frente a una amenaza cada vez más constante en la red, los ataques DDoS. Actualmente existen sistemas de detección de este tipo de amenazas, sin embargo, la parte innovadora de este trabajo reside en el estudio en profundidad de una red actualmente en auge pero muy poco investigada, la red Tor. Se ha hecho especial hincapié en evitar ataques de denegación de servicio en esta red, área donde la existencia se sistemas de detección es prácticamente nula. Para llevarlo a cabo se ha estudiado la estructura y funcionamiento de la misma y se ha elaborado un sistema de detección basado en anomalías dividido en dos partes, un modelado de datos basado en métricas para obtener valores de entropía y un modelo predictivo cuyo punto fuerte es la eficiencia sin dejar a un lado la calidad de predicción. La experimentación realizada tanto en la red usual como en la red Tor ha arrojado resultados muy satisfactorios obteniendo una alta tasa de verdaderos positivos y una baja tasa de falsos negativos todo ello sin un consumo exhaustivo de recursos del sistema, uno de los principales objetivos del trabajo realizado
The aim of this paper is to face DDoS attacks, a huge threat in the net. Nowadays there are several protection systems to confront them. Nevertheless this paper is made up of an innovative aspect. This is the deep investigation of the Tor net, which is increasing its popularity in the last years. This work gives emphasis to avoid DDoS attacks in this net where the existence of Protection system is almost null. To reach this goal the infrastructure and functioning of Tor have been deeply studied. In addition a detection system based in anomalies has been developed. It is divided in two different sections. The first one consist of a data model based in metrics to obtain entropy values. The objective of the second one is the development of an efficient prediction model that also takes care of the accuracy. The experimentation performed in the usual net, but also in the Tor net has shown really satisfactory results. The true positive rate obtained was really high compared with the false positive rate which has truly low rate.
The aim of this paper is to face DDoS attacks, a huge threat in the net. Nowadays there are several protection systems to confront them. Nevertheless this paper is made up of an innovative aspect. This is the deep investigation of the Tor net, which is increasing its popularity in the last years. This work gives emphasis to avoid DDoS attacks in this net where the existence of Protection system is almost null. To reach this goal the infrastructure and functioning of Tor have been deeply studied. In addition a detection system based in anomalies has been developed. It is divided in two different sections. The first one consist of a data model based in metrics to obtain entropy values. The objective of the second one is the development of an efficient prediction model that also takes care of the accuracy. The experimentation performed in the usual net, but also in the Tor net has shown really satisfactory results. The true positive rate obtained was really high compared with the false positive rate which has truly low rate.
Description
Trabajo de Fin de Grado en Ingeniería Informática (Universidad Complutense, Facultad de Informática, curso 2014/2015)