Research On The Optimization Of Online Education Platform Archi Tecture Based On Nginx

In recent years, the informatization of education has developed rapidly.And there are many excellent educational service products emerge comes into being. Among them, “cn901.com” is one of the most representative educational services website. However, the performance problems of the system have been gradually appeared, as the user base grow gradually. Such as, the response time is too long, the concurrency value is too small, and it has bad stability. So, the task of this paper is to solve these problems. We will optimized the system from an architectural perspective, in order to improve the system's concurrency, shorten the response time of the system, and implement the system's high availability, and improve the user experience ultimately.The main research results of this paper are as follows:(1) The design of the multi layer network architecture prioritization scheme. We extend and optimize the structure of this network based on the existing three tier architecture: first, make the application server be clustered, and then improve the system's service capacity; second, add load balancing layer for managing and scheduling the server clusters; at the data layer, add NoSQL database and search engine on the basis of SQL database, to achieve the separation of reading and writing.(2) The proposing of a dynamic load balancing algorithm based on request task classification.First, we make comparison of and analysis on a number of load balancing algorithms that commonly used. Based on that, we add the dynamic analysis of the remaining service capacity of each node in the server cluster. Meanwhile, all requests are divided into two types of CPU intensive tasks and IO intensive tasks. And then we proposed an improved algorithm of least connection,which is a dynamic load balancing algorithm based on request task classification. After that, we implemente this algorithm through further development on Nginx.The research results of this paper can improve the website concurrency significantly, and reduce the response time significantly under high concurrent environment. Moreover, the balancing algorithm proposed in this paper can distribute requests more balanced, in the case of a large number of CPU intensive task requests and IO intensive task requests access to the site at the same time.