Pleskanka M. Improving the quality of service in the content delivery network

The research is aimed at developing methods and algorithms for processing requests, that require significant computing resources, data caching, implementing load balancing and adaptive microservices deployment at CDN network points of presence, to improve the quality of user service in conditions of limited computing infrastructure. The paper considers the main principles and approaches related to the microservice architecture. A comprehensive analysis of the literature on dynamic load balancing and data caching, focusing on the current state of the strategies used, their advantages, disadvantages, implementation challenges, and their impact on the efficiency of CDN networks. An integral caching key is proposed, which is formed by connecting several parameters, and can be used for various types of network traffic. A method of processing requests that requires significant computing resources is proposed, which provides the possibility of distributed data processing at CDN network. A load balancing method in edge locations of the CDN network is improved, which takes into account the value of the integral caching key, the content availability on the edge server, and functioning state of available servers. A complex mathematical model of the queuing system, comprising three integrated subsystems, is presented to characterize the process of servicing requests in cloud services with dynamic resource allocation. Based on mathematical modeling, the quality of service parameters were assessed. For the first time, a method for adaptive microservice deployment at CDN network point of presence was developed. This method provides duplication of the business logic service from the origin server to edge servers, analyzes the parameters that trigger the quality of service in real time, and adaptively distributes requests based on the edge servers load. The practical significance of the results is that the method for processing static data requests in content delivery networks has been improved, which allows improve the quality of user service by reducing the response time for the end user by up to 9%, and the load of the origin server by 10%. The integrated use of the integral caching key, the method of processing and balancing traffic at CDN network points of presence allowed increase the caching efficiency by 30%, as well as reduce response time for end user requests by 40%. A method of adaptive microservices deployment for processing dynamic data in real-time mode at CDN networks points of presence has been developed, which makes it possible to actually provide the required quality of service in conditions of limited origin server resources.