Decomposition Method for Synthesizing the Computer System Architecture

Abstract

The paper considers the problem of synthesizing the architecture of a modern computer system (CS) for managing a complex distributed object. The main stages of the architecture synthesis are analyzed. The necessity of preliminary decomposition, structurization and formalization of the system is substantiated. The main stages of the hierarchical sequence of the decomposition process are given. The decomposition method for synthesizing the architecture of a computer system is developed on the basis of the stratification of the directed graph of CS implementation options. The simulation model of this process was developed. The modelling results showed the feasibility of using the developed method for synthesizing the architecture of a computer system with a large number of nodes. It is shown that the use of the decomposition method is unpractical when there are less than 60 nodes. Using more than 100 nodes is practically advantageous. This method, unlike the existing ones, enables getting results in a reasonable time when using more than 1000 nodes. These results are particularly significant in the context of operational reconfiguration of a system with a large number of nodes. Moreover, the more the number of nodes, the greater the advantage of the method of decomposition. Thus, while synthesizing the computer system architecture to control a complex distributed object at the stage of developing the relationship among the components, the method of decomposition should be used to reduce the time of the process.