Research On Key Technology Of Linux-based High-performance Parallel Agent System

In recent years, with the application of network users in the field of business increase,the network information security has become a focal point. In this way, agent technology arises at the historic moment. The benefits of the proxy server are numerous, so the number of users increased significantly. And many problems are generated in terms of performance.How to improve agent system in terms of performance, will directly affect the response speed of the user.This paper based on the high performance of the agent system as the research direction,Key research agency service single node in the cluster performance improvements, as well as in the cluster, load balancing efficiency. In view of the performance improvement of single node, put forward a memory allocation solution. For the cluster between each processing, the paper puts forward an improved load balancing algorithm.First of all, the dynamic memory management. This paper proposes a realistically consider dynamic memory management scheme. This scheme is designed for mass connection management network service program. It compared with the classical dynamic memory management scheme,simplifies the distribution and recovery algorithm, and get a great improve in terms of performance. And increases the dynamic tracing of allocation and release, in the program memory leaks, can immediately locate the leak location and modified.Second, load balancing algorithm between the server cluster. The algorithm is based on the combination of the predict state of the network and the current state of the node. Judge proxy node status first, and then predict the network status, integrated to calculate the feedback of the current node to send to the load balance node. Algorithm fully considered the differences between the various service node, can cause a high performance service node are fully utilized. improve response speed in high limit.Finally, according to dynamic memory management method and the load distribution algorithm, design a server cluster system, the various modules of this system.