OPTIMIZATION OF LOGGING IN INFORMATION SYSTEMS BASED ON A MATHEMATICAL MODEL OF BUFFER MANAGEMENT IN THE PRODUCER–CONSUMER PATTERN

Abstract

The article presents a comprehensive approach to optimizing logging processes in information systems by utilizing a mathematical model for buffer management within the widely used Producer–Consumer interaction pattern. The study addresses the challenges associated with high-frequency logging, such as data loss, increased latency, and resource overload, which are common in distributed and high-load environments. To mitigate these issues, a dynamic buffer management model is proposed that adaptively regulates the interaction between log-generating components (producers) and log-handling or storage subsystems (consumers). The model takes into account critical system parameters, including buffer size, event generation rate, and consumer processing speed. It enables dynamic adjustment of logging strategies depending on the current load and buffer state.

The research includes the formalization of the proposed model using discrete-time mathematics, with particular attention to queue theory and finite-buffer constraints. Simulation experiments conducted in the study demonstrate that the model significantly reduces log loss, optimizes system responsiveness, and ensures stable operation of logging mechanisms even under extreme workloads. The findings suggest that implementing such a model contributes to the design of resilient auditing and monitoring subsystems, especially in cybersecurity-sensitive and mission-critical infrastructures. The approach can be integrated into various architectural layers of modern information systems, improving reliability, maintainability, and traceability of operations in real time. Recommendations for practical implementation and possible extensions of the model for adaptive load prediction are also provided.