The DIDAPPER (distributed intrusion detector with apperception) system presented in this paper is a distributed intrusion detector with apperception. The distributed architecture, the apperception ability and the sharing of knowledge are evident characteristics of the DIDAPPER. This paper focuses on the apperception ability of DIDAPPER. Traffic specimens and IP traps are DIDAPPER's new concepts, which can capture and recognize abnormal traffics and are suitable for monitoring the large scale network attacks. The other aspect of DIDAPPER's apperception ability comes from the neural network algorithm. The BP neural network with learning ability has been applied to traffic analysis, and shows good effect on the recognition of traffic patterns.