Problem analysis and improvement study of low-performance web application: An example of Student Information System course selection module

All of the universities have student automation systems in which all of the student's education processes are monitored. It is known that in the course selection week at the start of the semester, these systems or underlying structure usually cannot handle the high traffic and service outages are experienced. In this study, the student information system used in Bilecik Şeyh Edebali University is re-analyzed and service disruption eliminated by optimizing software architecture, SQL query structure and server configuration. MVC architecture is used instead of Web Form architecture in the re-designed system. Design of the front-end is implemented with Bootstrap and pages are designed to be responsive for different window, screen sizes. While analyzing the old system it is seen that most calculations are done in C# back-end and these calculations are moved to client side by using JavaScript. This allowed the calculations to be performed by the student computers instead of web server. Complex database queries that are not depended on each other's results for these calculations are queried in parallel with multi-threading method in order to achieve speed. Typical system delay for an operation to this eliminated service outage by reducing waiting time in the queue for IIS. In addition to software improvements, a load balancer application is employed in order to distribute the client traffic to four servers with a balancer server in front of them instead of running the software in a single server. Even if high client traffic leads to problems, this method allows the system to operate without service interruption by adding additional worker servers in the load-balancing configuration. The solutions described here for eliminating the service outage for the student automation system can be applied to all automation systems that suffer from same problem due to high volumes of traffic. Key Words: MVC; Multi-thread; Load Balance; Software Performance; Automation