This document describes a proposed network intrusion detection system using attack behavior classification. The system aims to maximize recognition of network attacks by embedding their temporal behavior patterns into a neural network structure. It captures packets in real time using an engine that preprocesses data and sends it to modules for pattern recognition, classification, and generating alerts. The system was tested in a real environment and showed ability to detect attacks. It aims to address limitations of existing systems like constant monitoring overhead and inability to distinguish threats from normal traffic.