MSMykola Sliepushkov

Mykola Sliepushkov

*This is an algorithmically generated author record based on publicly available information.

Mykola Sliepushkov has a total of 6 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    MODIFIED SOFTWARE DEPLOYMENT ALGORITHM USING MULTI-THREADING

    by Nataliia Khatsko, Mykola Sliepushkov, Kyrylo Khatsko, Yevhenii Shebanov

    The article presents a modified deployment algorithm for software systems using multithreading in AWS CodeBuild, aimed at optimizing build time and reducing computational resource costs in cloud environments. The key stages of the build process, including parallel test execution, task allocation analysis, and resource management, were modeled using finite automata, timed automata, and Petri nets. Particular attention was given to identifying and addressing the limitations of AWS CodeBuild's standard parallelization mechanisms, which can lead to inefficient resource utilization and extended build durations. The study revealed that AWS CodeBuild's default mechanisms are not always capable of optimally leveraging system resources, especially when handling large software projects with numerous dependencies. To overcome these limitations, the use of Python's multithreading capabilities was proposed as a convenient tool for extending the platform's base functionality. The proposed approach enabled flexible thread management and task distribution at the user scenario level, significantly reducing overall build time. Experimental results demonstrated substantial reductions in build execution time compared to the default AWS CodeBuild settings, confirming the effectiveness of the proposed algorithm in enhancing performance and ensuring high scalability for build processes in cloud environments. The developed algorithm is particularly relevant for large software projects requiring frequent iterative builds and testing. The findings can be utilized to improve automated deployment processes and computational resource management in cloud ecosystems.

  • Open Access

    RESEARCH OF THE MANAGEMENT OF SOFTWARE SYSTEM DEPLOYMENT USING THE RESOURCES OF CLOUD PROVIDERS

    by Andrii Pashniev, Mykola Sliepushkov, Denys Hurt, Iryna Liutenko

    The main stages of the process of deploying a software system using the resources of cloud providers are analyzed. The models of possible options for managing the deployment of a software system using the BPMN notation are developed. Based on the modeling, the advantages and disadvantages of non-automated, automated, and automatic management of the deployment process are identified. A model of automatic deployment of a software system using the resources of cloud providers' data centers in IDEF0 notation has been developed and decomposed, which allowed to study the functional interaction of individual stages of the software system deployment process. Based on the developed contextual and decomposition IDEF0 diagrams, within the framework of modeling the software system deployment process, an analysis of the functional interaction of the Build and Deploy stages was carried out, which showed that these stages have the greatest functional load and require more detailed research in order to find ways to optimize them. With this purpose, were developed and analyzed IDEF0 and DFD decomposition diagrams that model the functional interaction and data flows between the components of the Build and Deploy subprocesses. Analysis of the functional interaction and data flows generated and transferred between the individual components of the Build and Deploy subprocesses revealed an important aspect, which is that each time the source code is compiled, and the container is created, data must be loaded from external libraries. This, in turn, leads to a significant increase in external traffic, which affects the speed of deployment of software systems using the resources of cloud providers' data centers. As a possible way to optimize the Build and Deploy subprocesses, was proposed to create a cache to save the downloaded data from external libraries for reuse, as well as implement control over queries to external libraries and the data caching process, which will minimize the time spent on transferring data from external libraries during automatic deployment.