Agent-based Intellectualization of Continuous Software Integration and Resource Configuration for Distributed Computing

The need to carry out large-scale scientific experiments in distributed environments is growing from year to year. Currently, development tools of scientific packages, as a rule, do not fully support continuous integration of applied software. This negatively affects the software quality and increases both the time for its development and the time for carrying out experiments. Within this paper, we introduce a new set of methods to use the best practice of continuous integration of system and applied software in developing distributed applied software packages based on a modular approach. The proposed approach is implemented in the Orlando Tools framework. It distinguishes two levels of continuous integration. At the first level, we solve the problems of integrating the Orlando Tools means with the GitLab system and automating the module development process. The second level, associated with problem-solving in combining modules within a package and placing it on the environment resources, is fully implemented using Orlando Tools. Thus, the framework provides automation in developing program code and its building and testing at the module level. In addition, it ensures the package deployment on the environment resources with its subsequent testing as a whole. Moreover, Orlando Tools provides the necessary configuration and customization of environment resources. A distinctive feature of the proposed approach is intellectualizing continuous software integration and resource configuration using multi-agent technologies. The practical significance of the study results consists in the decrease in the number of errors and failures in the development and use of packages. This reduces the time required to carry out experiments and increases the efficiency of using resources.