| Abstract: |
Automated generation of code, architectural solutions and test scenarios speeds up the development processes, but places a significant burden on the verification, testing and validation stages. Moreover, the expenditure of resources on these processes is justified only if the developed or generated software meets the initial requirements, constraints and context of use. In the absence of systematic documentation, formal evaluation models, and integrated analysis, it is difficult for even experienced architects and engineers to predict the consequences of using generative tools, let alone managers and business analysts who evaluate the project at the initial stage. In such conditions, software functional state (SFS) analytics tools become particularly relevant throughout the entire SDLC. They will not be able to integrate disparate indicators of quality, security, performance, reliability, and contextual factors into a single system state model, providing justification for decisions on resource allocation, risk assessment, technical debt management, and strategic product development planning. The publication discusses the principles of designing such analytics tools based on an analysis of existing models, methods, and approaches to assessing SFS based on modeling the most common scenarios throughout the SDLC. Special attention is paid to the structuring of analytics by layers: descriptive, diagnostic, predictive, prescriptive, comparative and contextually adaptive. Such a hybrid multilayer model enables not only the recording of the current SFS but also the identification of the causes of transitions between states, the prediction of degradation or the emergence of vulnerabilities, the formulation of recommendations for optimal management actions, and the evaluation of alternative development strategies. The paper summarizes the SDLC model, taking into account modern methodologies and identifies typical metrics, analytics goals, and expected results of its application. The proposed analytics tools were developed based on the scenarios that are most typical in the IT industry, which allowed us to estimate their feasibility, adaptability and potential for early risk detection and decision-making support at the stages of requirements formation and system design. [ABSTRACT FROM AUTHOR] |