The Mathematical Model for Assessing the Reliability of Multiprocessor Computing Systems Functioning

For the widespread use of information technology in the organization’s business process, which will be aimed at optimizing work and leading to increased productivity and profitability, quality software is needed. Consequently, the design and production of new software (software) requires an accurate analysis of its technical characteristics and, on this basis, will remain one of the pressing tasks in the field of information technology. Therefore, the article discusses an approach for assessing and improving the basic parameters of effective software operation. Reliability, to ensure the required performance, is the main criterion of operation, since it is the ability of a software product to reliably perform specified functions under specified conditions for the required period of time with a sufficiently high probability. The problem of software reliability deserves more and more attention due to the continuous complication of the systems being created, the increase in the range of tasks assigned to them, and, as a conclusion, a significant increase in the complexity and volume of software. New versions are used for those software modules that may experience software failures. To implement the proposed approach, a mathematical model for assessing software reliability is provided. Formulas are presented that are used to calculate the complex reliability parameters of the system under consideration. Relevant examples are shown. For this purpose, the Markov service model was used, that is, the study of queuing systems using a Markov process, which has a discrete set of states. The process of functioning of a multiprocessor computing complex consisting of identical processors is represented by a closed queuing system with waiting.