GL-Model, Representing Emergence Of Dangerous State In Multiprocessor Management System
Authors: Rabah AlShboul, Vitaliy A. Romankevich
Number of views: 242
Graph-logic models (GL-models) are models of fault-tolerant multiprocessor systems’ (FTMS) behavior in failure flow. Basis of the model is a graph which arcs are marked by some Boolean functions and the arguments of these functions are system processors states. GL-models are tools that allow efficiency increasing of reliability calculation of FTMS. The reliability calculation is based on the method of statistic experiments with GL-models.We propose the idea and method of creating GL-model, which answers the question if the multiprocessor complex object management system is technically (functionally) safe in the case when the failure of certain components in the system occurs. In other words, if there is a dangerous condition in the system, that is a state of the management system, in which at least one management function, that causes a dangerous condition, is not performed. The model can be used to conduct statistical experiments to determine the probability of transition to a dangerous state in case of system’s degradation.