The Achieving Of Dynamic Load Balancing System In Distributed Server Cluster

In recent years, with the rapid development of mobile Internet, the traditional services of telecom operators (such as voice, SMS and MMS) have been squeezed by the mobile Internet business. The living space of telecom operators becomes smaller and smaller, it is the time to find a way to strike a place in the mobile Internet market, and PaaS cloud technology is the best choice. PaaS can integrate operators’IDC server and pipeline resources, gather the developing ability of various types of applications. In this way, they can reach the goal of reducing operating and maintenance costs, enriching business types and diversifying revenue models.This thesis focuses on the realization of the underlying PaaS cloud technology-Dynamic load balancing system for distributed clusters. The technologies of distributed server cluster and load balancing were researched and analyzed. Then, based on the preliminary investigation of the commercialized PaaS cloud platform products, a design of dynamic load balancing system architecture and a dynamic load balancing scheduling method were proposed. They form the load balancing system. The proposed method is based on weighted round robin algorithm and particle swarm optimization algorithm, and the system can realize the dynamic load balancing of application accessing requests and dynamically allocate the resources for applications according to the actual needs of the server resources. Finally, in laboratory experiment, a dynamic load balancing verification system for distributed server clusters was built to verify the feasibility of the proposed design.In summary, this thesis gives a set of implementation solutions for each function node of the dynamic load balancing system for distributed cluster. These solutions can be used for reference in future commercial building and standardization.