Functional stability of information networks in the availability of limited a priori information about reliability
In the article, based on the use of the results of evaluating of linear and fractional-linear functionals, sets of calculated ratios are obtained to determine the guaranteed (largest and smallest) values of basic of the reliability indicators of information systems in the presence of limited a priori information about the distributions of determining random variables which are represented by the initial moments. Such calculations at the design stage of complex information systems are predictable, because they serve as a grounding for the expected indicators of reliability of information systems that do not yet exist in real life. The peculiarity of the considered information systems is that they are used not continuously, but sporadically. In this case they perform tasks that come at random times. In such systems, the replenished time reserve provided in the system itself is used along with the random time reserve due to the nature of the task (this time reserve is determined from the moment of failure of the element or information system until the receipt of the task). Examples of sporadic information systems are VANET car networks, which have high latency requirements, real-time data centers, and communication systems. The main purpose of these calculations is as follows: a comparative analysis of various design (scheme) options for information systems at the design stage for a reasonable choice of the general structural scheme, methods of redundancy, methods of control and maintenance; a predicted assessment of the reliability of the system is given to substantiate the guaranteed conclusion that the designed information system can be manufactured according to the requirements that meet the requirements of functional stability; an approximate assessment of the reliability of information systems at the stage of testing a test sample for a reasonable determination of the terms of maintenance of systems in conditions of a priori uncertainty.