RESEARCH ON THE EFFICIENCY OF PROCESSING PARALLEL REQUESTS BY AMAZON WEB SERVICES USING CONTAINERS

• Published in: Електроніка та інформаційні технологіі Vol. 29; no. 29; pp. 47 - 56
• Main Authors: Sigunov, Oleg, Demkiv, Lidiia
• Format: Journal Article
• Language: English
• Published: Ivan Franko National University of Lviv 01.03.2025
• Subjects: aws ecs; aws eks; cloud services; containers; gatling; instances; testing scenarios
• ISSN: 2224-087X; 2224-0888; 2224-0888
• DOI: 10.30970/eli.29.5

Background. The article explores the efficiency of a web application deployed in AWS cloud services using containerization based on AWS ECS and AWS EKS. Particular attention is paid to analyzing system performance under high load conditions typical for online services that provide real-time sports competition updates. The ability of the application to handle a large number of user requests simultaneously is assessed, which is critically important for ensuring proper operation in real-time mode. Methods and tools. The research methodology is based on load testing using modern tools for generating requests and collecting metrics. A series of experiments were conducted to model different usage scenarios, including peak loads, stable operating conditions, and situations with sudden traffic changes. Key performance parameters were analyzed: server request processing speed, system stability under increasing load, and the proportion of unsuccessfully processed requests. Results and discussion. The experimental results demonstrated that AWS EKS provides the best ratio of performance, stability, and cost for deploying similar solutions. The use of Kubernetes clusters allows for flexible infrastructure scaling, adapting to variable loads without significant losses in speed. While AWS ECS ensures sufficient performance, it proved to be less flexible compared to AWS EKS, particularly in cases of dynamic load changes. Conclusions. Thus, the study confirmed the feasibility of using AWS EKS as the primary platform for containerizing a high-load web application. The obtained results may be useful for developers and cloud solution architects working with real-time systems requiring high performance, flexibility, and reliability. Additionally, the presented data may serve as a basis for further research in optimizing cloud resources and improving the efficiency of web services.