THE PROCEDURE FOR ENSURING FAULT TOLERANCE OF A COMPUTER SYSTEM BASED ON THE USE OF MODULAR ARITHMETIC

The article proposes for consideration the process of functioning of a computer system (CS) in modular arithmetic (MA) based on the use of the principle of fault tolerance. In MA, the fault tolerance of the CS is ensured by taking into account the main properties of this modular numbering system, namely: independence, equality and low-bitness, the totality of which forms a modular code structure. In order to facilitate and visualize the study of the functioning of the computer system in the MA, the article introduces an n-bit binary code that reflects the set of signs of the operational state of the computing tracts of the CS. With the help of this code, an analysis of various modes of fault tolerance of CS in MA was carried out. The fault-tolerance property provides the possibility of performing specified computing functions after the occurrence of failures, both due to a decrease in the permissible limits of any indicators of the quality of functioning, and without deterioration of the indicators of the quality of the functioning of the CS. The article considers the mode of fault tolerance of CS in MA, which is implemented by the method of replacement. The essence of which is as follows, if the reserve is exhausted (reserve computing path), then the failure-tolerance mode of the CS in the MA is implemented by the method of gradual degradation. CS in MA can be considered as a multi-microprocessor computing system with dynamic redundancy and subsequent gradual degradation of functionality in the event of failures. The structure of the functioning of the fault-tolerant CS in MA has been developed. Examples of CS fault tolerance for a specific set of MA modules are given. The comparative analysis of the research results carried out in the article showed the following: with fault-tolerant functioning in the modes of replacement or replacement with subsequent gradual degradation, the CS in MA is more effective than the similar CS in the binary positional numbering system. Using MA as a counting system allows you to create a CS structure similar to the structure of a reserved computing system in a binary positional numbering system. This circumstance, along with other positive aspects of the use of modular coding, caused by the influence of the main features of MA on the structure and principles of operation of the CS, allows to consider MA as a powerful tool for increasing fault tolerance.

Yanko, A.

active fault tolerance

Національний університет «Полтавська політехніка імені Юрія Кондратюка»

info:eu-repo/semantics/article

application/pdf

https://journals.nupp.edu.ua/sunz/article/view/3168

Control, Navigation and Communication Systems. Academic Journal; Vol. 4 No. 74 (2023): Control, Navigation and Communication Systems; 125-128

uk

https://journals.nupp.edu.ua/sunz/article/view/3168/2574

Авторське право (c) 2023 A. Yanko, V. Krasnobayev, R. Liubchenko, P. Sabelnikova