An adaptive security end-system model based on active defense

In order to enhance the security of Web service as well as to support the differentiated service, the paper puts forward an adaptive model, which is based on active defense strategy. It uses a network prediction model on base of self-similar sequence theory to estimate current slates of transfer servers in end-system, then classifies them with fuzzy-C-means clustering method. Finally, we bring forward a load balancing algorithm that is founded on the classification results and grading users' requests. In that way, it optimizes the allocation of system resources and ensures data security and reliability. Moreover, this model is not dependent on the original operating system and is transplantable and extensible