Diagnostic Model for a Distributed Computer System in Real Time
- PDF / 597,219 Bytes
- 11 Pages / 612 x 792 pts (letter) Page_size
- 59 Downloads / 219 Views
UTER METHODS
Diagnostic Model for a Distributed Computer System in Real Time A. M. Gruzlikova,*, N. V. Kolesova,**, E. V. Lukoyanova,***, and M. V. Tolmachevaa,**** a
AO Concern TsNII Elektropribor, St. Petersburg, Russia *e-mail: [email protected] **e-mail: [email protected] ***e-mail: [email protected] ****e-mail: [email protected]
Received October 15, 2019; revised May 19, 2020; accepted May 25, 2020
Abstract—An approach to test the diagnostics of distributed computing systems based on the introduction of redundancy in the system in order to simplify the test construction process and the diagnostic process is investigated. The introduced redundancy is a diagnostic model of the system, which runs in parallel with the main software of the system. For this model, sufficient conditions for controllability and observability are formulated. DOI: 10.1134/S1064230720050068
INTRODUCTION The problems of diagnosis and fault tolerance occupy an important place in the process of developing control systems and information processing. Despite the fact that research in this area has been carried out for several decades and their results are reflected in numerous publications, in practice new tasks that require in-depth study are continuously being forward. At the same time, one of the most significant factors prompting new research is the continuous increase in the complexity of systems and the multiplicity of causes of the violations, including hardware failures and malfunctions, as well as software errors. The problem of the high dimensionality of the problem is overcome in practice by its decomposition based on a hierarchical approach, in which the system’s components are placed according to complexity levels. Further, for each level, its own diagnostic tools (DTs) are developed that detect violations in the information connections between the components. The basis of the solutions used in practice in the synthesis of DTs are the techniques of functional and test diagnosis [1–6]. As the mathematical model of the diagnosed object, a Petri net, a dynamic system, and a finite state machine are usually used. It is known that the solution to the diagnostic problem is always related to the use of hardware, information, or temporary redundancy in the system. The DTs themselves are an example of such redundancy. In addition, prior to the synthesis of DTs, redundancy can be introduced into the system in order to simplify the diagnosis process. In particular, these may be steps to increase the number of inputs or outputs of the system, introduction of special analysis tools in the system, etc. The introduced redundancy can formally be attributed both to the system and to the DTs, which is insignificant. The approach proposed earlier [7–10] and developed by the authors in this study, which involves the introduction of redundancy to simplify the diagnosis, belongs to this direction. This article discusses real-time distributed computing systems (DCS), which form the core of modern information management and processing sys
Data Loading...
