Research And Implementation Of Web Server Performance Optimization Based On The Lamp Architecture

This paper studies the working principle of LAMP architecture Web server and existing Web server performance optimization techniques. We research the major problems and solutions existing in the LAMP architecture Web server optimization process. There are three major problems,1) How to conduct the performance test, so that the test results can better and more accurately reflect the performance of the server;2) Because the LAMP architecture has powerful configuration, without changing hardware environment condition, how to make servers to achieve high performance quickly by modifying the configuration parameters;3) Cluster technology is a commonly used method to optimize Web server performance. How can we use clustering technique to optimize the LAMP architecture Web server in a better way. In this paper, we obtain the following results based on the above issues:1) We propose a performance testing model of the LAMP architecture Web server. We improved the classical Web performance testing model (WPTM) by increasing the user behavior model. We design testing cases based on user behavior model, then we use Web performance testing tool testing the performance of Web server. User behavior model mainly include the distribution of different types user operations and user thinking time. Traditionally our mathematical modeling based on statistics. In this paper, we model user behavior by analyzing the apache logs.2) We propose a parameter optimization algorithm of the LAMP architecture Web server. In LAMP architecture Web server, there are many parameters that can be adjusted and optimized. The traditional parameter optimization algorithms include enumeration algorithm, and greedy algorithm. MaxClients parameter of apache also has a relatively rough estimation algorithm. However, this estimation algorithm does not take into account the resources consumed by the operating system. Considering this respect, we propose an improved estimating MaxClients parameter algorithm. Parameter optimization problems is a kind of combinatorial optimization problem. This paper proposes a pararmeter optimization algorithm which combing climbing search algorithm and the improved MaxClients parameter estimating method.3) Software structure of LAMP architecture Web server is a kind of hierarchical structure. When building the LAMP architecture Web servers cluster, we should think about the relationships between the software. There are two ways to extend the LAMP architecture Web cluster. One is increasing the number of nodes in the cluster, the other is increasing the performance of the server cluster nodes. Under different user load conditions, we analyze the advantages and disadvantages of the two methods. We proposed a based user loading Web server cluster scaling selection strategy.