Load balancing optimization method based on HTTP request classification

The invention discloses a load balancing optimization method based on HTTP request classification. The method comprises the following steps: receiving an HTTP request from a client; determining the type of the HTTP request by analyzing the Content-Type content in the HTTP request message; setting an initial weight value for each server based on the static performance of each server, and calculating and sequencing the current weight value of each server according to the determined type of the HTTP request; and selecting the server with the highest current weight value to establish connection with the server. The invention further discloses a load balancing optimization system based on HTTP request classification. The system comprises a request classification module, a server selection module, a request distribution module and a downtime prevention module. The client HTTP request classification and the dynamic load balancing strategy are combined to select the proper server for processing the HTTP request, so t