Methods for quality assessment of standardized software time indicators

Abstract

Time parameters of software quality indicators are among the most important when assessing the quality of software products, as they determine performance, stability and resource utilization efficiency. In the article time indicators of software product quality are researched in accordance with international standards ISO/IEC 25022 and ISO/IEC 25023. These standards provide a list of software quality indicators, but do not contain any methods or techniques for determining them. In particular, there are no any measurement techniques for time indicators of software quality. The purpose of this research is study and systematize time indicators of the software products quality and develop methods for their classification and evaluation on a time scale. The feasibility of forming a clear classification of time metrics is substantiated, which covers different time scales – from immediate system reactions to long-term effects of exploitation. The method of structuring indicators into four time ranges (seconds, minutes/hours, days, months), which provides a complex assessment of performance, reliability and stability of software products. An application has been developed that implements the measurement of basic time indicators, in particular, mean response time, median, 95-percentile and dispersion, based on multiple HTTP requests using high-precision timer. The experiment with SLA threshold equals to 150 ms confirms a practical efficiency of the proposed model and it eligibility to industry standards. The obtained results show high performance and stability of the system, that indicates research theoretical statements is relevant. The research emphasizes, that time indicators are a universal indicator of software quality, providing a basis for standardized performance assessment for different types of information system.