Mathematical model of reliability of information processing computer appliances for real-time control systems

One of the main characteristics of computer appliances for processing real-time information is reliability. The reliability of software is understood as the property of this software to perform specified functions, maintaining its characteristics within the established limits under certain operating conditions. Software reliability is determined by its reliability and recoverability. Reliability of software is a property to maintain its performance when using it for processing information in the information system. The reliability of the software is estimated by the probability of its operation without failures under certain environmental conditions during a given observation period. The development of real-time systems requires a large amount of resources for design and testing. One of the solutions to this problem is mathematical modeling of computer appliances. This allows more flexible design of real-time systems with the specified reliability, taking into account the limitations on price and development time, and also opens the possibility of more flexible optimization of computer appliances for real-time control systems. To develop a mathematical model of the reliability of computer appliance for real-time systems, it is necessary to take into account the provision of a given level of reliability, with reasonable development costs. There are many methods for improving software reliability, but the most promising and effective methods are redundancy, which is achieved using N-version programming. To increase the reliability of the hardware of the computer appliance, it is also necessary to use redundancy and redundancy, which includes multiprocessor and provision of different buses and independent RAM. This paper discusses existing approaches to improving the reliability of hardware and software, proposes a model of reliability of a computer appliance, which is understood as the product of the probability of failure-free operation of hardware and the probability of error-free operation of software. In addition, new formulas are proposed for the steady state probabilities of the hardware states of a multiprocessor computer appliance with heterogeneous processors, which give the same result as the existing ones, but require fewer computations. The paper concludes with a question about the possibility of optimizing the reliability of computer appliances based on the developed model, and indicates optimization methods that can be used to solve this problem.