Um sistema P2P para detecção de anomalias de rede

Abstract

Precise and efficient monitoring is vital to ensure that a network works according to the intended behavior, as well as quickly acting to the problems found. The task of monitoring a network becomes complex with increasing network size and heterogeneity. The available network monitoring and management solutions are not only costly but also difficult to use, configure and maintain. This work aims to continue the development of a P2P system to detect network anomalies to help network administrators. The system should be easily used by ISP network administrators. Initially, state-of-the-art research will be presented, where various technologies related to the objective of the dissertation will be highlighted. Multiple requirements of monitorization systems will be exposed and there will be a description of the applicability of a P2P system in a monitorization system. Afterward, all the processes of creation of the system and the existing entities will be described as well as their communication capabilities. All the ways an administrator can interact with the system will also be presented. Then, the process of definition of a metalanguage will be exposed, to allow the administrators to configure and pre-program the network in an effective and varied manner. The process of optimization of the monitoring system will be described, to reduce the traffic of the P2P network. It will be also described how the system became fault-tolerant, recovering its state if any entity has a problem. Subsequently, the implementation of each one of the developed mechanisms and the architecture of the system will be exposed, where it’s explained which technologies were used and justified some paths chosen to achieve the dissertation objectives. The system will be tested in a network emulator and the resulting data from the created mechanisms will be analyzed to validate the correct behavior of what was developed.