This paper uses Web Service technology to design and develop a dynamic detection platform. The platform focuses on the detection and analysis of anomaly traffic for large network flow architecture. This article will define each type of anomaly detection methods as a service that is implemented by Web Service technology. The framework has the flexibility that allows administrators to adjust the resources according to the circumstance variation by enabling or disabling different detection services, and add new services to accommodate a variety of new attacks. To deal with the large network flow environment, we use virtualization technology to manage the traffic database through a distributed system infrastructure. Combined with the library of read/writing splitting mechanism, the system can distribute queries of Netflow data to different agents to share the access burden of a single database and improve the overall analysis speed and efficiency.