U-D method: An integration method for enhancing intrusion detection

Intrusion Detection has shown great potential in network security research. Most existing intrusion detection methods treat all data in the network as a whole. However, in reality, data in the network could be divided into two categories: upload data and download data. When intrusion takes place, these two types of dataflow may have different characters. Based on this discovery, we proposed a novel intrusion detection method (U-D method) taking both upload and download data into consideration. With the enhanced separately analysis method, we could figure out intrusion clues more effectively and efficiently. We wonder the relationships between these data might contain some instinct clue for discovering important intrusions. Experiment results demonstrate the effectiveness of our approach.