Intelligent Software Service Configuration Technology Based on Association Mining

Abstract

Association relationship and types between software service configuration parameters determine that the configuration items must meet the corresponding rules and constraints in parameter value and setting order. However, the distribution of software services and the architecture based on the loose coupling of heterogeneous components make the association of configuration parameters more concealed. The correlation and constraints of configuration parameters are crucial to ensure system configuration. When deploying, migrating, and updating systems, violating constraints will lead to configuration errors, resulting in system failures. To minimize the loss caused by configuration errors, association rules between configuration items can be mined to avoid potential incorrect settings. This paper proposes an intelligent software service configuration technology based on association mining by analyzing configuration items. The sample data of the configuration file is extracted from the open source code base, and the configuration file is parsed to obtain the configuration item name and value. Then mine the association rule of each configuration item pair based on the configuration item name, value, and type. According to the mined configuration constraints, configuration item error detection, cascade update and value recommendation are performed. Finally, the proposed method is tested and evaluated on six typical open source software, and the experimental results prove the accuracy and effectiveness of the method in this paper.