MICROSERVICE ARCHITECTURE FOR BUILDING HIGH-AVAILABILITY DISTRIBUTED AUTOMATED COMPUTING SYSTEM IN A CLOUD INFRASTRUCTURE

Abstract

The article is devoted to the research and development of a highly available distributed automated computing system by iterative algorithms based on the microservice architecture in a cloud infrastructure. The subject of the research is the practical foundations of building high-availability automated computing systems based on microservice architecture in a cloud-based distributed infrastructure. The purpose of the article is to develop and to substantiate practical recommendations for the formation of the infrastructure of a high-availability automated computing system based on the microservice architecture, the choice of its constituent elements and their components. The task of the work: to identify the necessary structural elements of a microservice automated computing system and to analyze the constituent components and functional load for each of them, set specific tasks for building each of them and justify the choice of tools for their creation. In the course of the research, methods of system analysis were used to decompose a complex system into elements and each element into functional components, and tools: information technologies Apache Kafka, Kafkacat, Wolfram Mathematica, nginx, Lumen, Telegram, Dropbox, and MySQL. As a result of the study, it was found that the system infrastructure should consist of: fault-tolerant interservice transport, a high-availability computing microservice, and communication microservices with end customers, which save or process the results. For each of them, recommendations are provided regarding the formation and selection of implementation tools. According to the recommendations, one variant of implementation of such system has been developed, the principles of its operation are shown and the results are presented. It has been proven that when using a Kafka queue it is efficient to publish batches of results rather than one at a time, which results to significant overhead on queue servers and data latency for its clients. Recommendations are given on the implementation of the CI/CD system to build a continuous cycle of adding and improving microservices. Conclusions. Practical foundations have been developed for the implementation of high availability distributed automated computing systems based on microservice architecture in a cloud infrastructure. The flexibility in processing the results of such a system is shown due to the possibility of adding microservices and using third-party analytical applications that support connection to the Kafka queue. The economic benefit of using the described system is shown. Future ways of its improvement are given.