SOFTWARE SYSTEM ARCHITECTURE DEVELOPMENT FOR INTELLIGENT ANALYSIS OF WEB APPLICATION PERFORMANCE METRICS
Liubov Oleshchenko, P.V. Burchak · Scientific notes of Taurida National V I Vernadsky University Series Technical Sciences · 2024
Today the number of web applications that process large amounts of data is increasing, which creates new challenges for developers and users.Web applications that handle big data are becoming an essential part of business, government, and everyday life, enhancing efficiency, accuracy, and speed of decision-making.This brings numerous problems, such as ensuring data security and confidentiality, efficient processing and storage of large volumes of information, and the need for continuous monitoring and optimization of web application performance.Resource management and energy efficiency are becoming particularly relevant in the context of energy savings.Solving these problems requires the use of the latest technologies, such as intelligent monitoring and analysis systems, which help ensure the stable and efficient operation of web applications in conditions of constantly increasing data volumes and task complexity.Addressing web application performance issues through intelligent monitoring systems will enable developers to quickly and accurately identify bottlenecks and optimize code, which in turn ensures a better user experience and reduces maintenance costs.Unresolved issues include the complexity of integrating new technologies into existing systems, the need to consider various factors affecting performance, and ensuring a high level of security and data confidentiality during monitoring.The article analyzes existing software solutions such as Google Lighthouse, Apache JMeter, New Relic, Dynatrace, GTmetrix, Pingdom, AppDynamics, WebPage Test, Sentry, and LoadRunner, their functional capabilities, main advantages, and disadvantages.It examines the possibilities of using machine learning and artificial intelligence technologies in the considered software systems.Based on the analysis, a software system architecture is proposed for analyzing the performance of web applications written in JavaScript, which allows for the collection of numerical data on the factors affecting the performance of the web application, performing regression analysis to determine the assessment of the influence of factors, clustering and classification of the processed data for the correctness of the allocation of recommendations for developers, which must be used in order to improve the performance of the web application and reduce the load on the web server.According to the conducted research, the use of machine learning methods in software systems for web application performance analytics can increase the performance of web applications by an average of 20%.