The 21st century is the information age. Because of the network infrastructure and technology has great development, network traffic increases very fast. For the network security, the network service provider should monitor traffic in a large scale network. They are challenged by handling a huge amount of traffic data for processing and management. The traditional intrusion detection systems (IDSs) are not enough to handle such Big Data efficiently. A flexible, effective network IDS is necessary to meet this case. In recent years, as the development of cloud computing technique, the cloud platform Hadoop which is the critical big data solution can provide an opportunity to process massive data. Mapreduce framework of Hadoop provides programmers the ability to produce parallel distributed programs more easily. Therefore, it is rational to apply this parallel computing architecture for the large traffic data monitor and analysis application. We created an IDS system based on support vector machines (SVM), and integrated it into Hadoop. Our goal is to run the detection function in parallel so that we can accelerate the analysis process. We conducted a series of experiment for the system evaluation. According to the experiment result, the proposed system has an impressive performance in the big data traffic analysis environment.