Multi-level Method of Behavioral Online Testing of Distributed Information Systems

Abstract

Systems for automated technical diagnostics of promising complex distributed information systems, often with critical properties of application, as well as actively added agent properties of autonomy, mobility, intelligence, and cooperativity, should ensure their complete real-time testing. Moreover, the online testing should be carried out both for the static hardware-software and information environment of the system location, and for the dynamic aggregate of information-check flows corresponding to the main functions and scenarios of its work. This paper discusses a multi-level method of behavioral online testing, based on a multi-level model for check of behavior of hierarchical Petri nets. The check model allows us to present both the active processes of the basic functioning of the static environment for placing distributed information systems and dynamic information and control flows in it, as well as passive processes of behavioral online testing of the environment and flows. The proposed online testing method differs in the background hierarchical organization of the temporal development of processes and check flows with evolutionary optimization based on multi-agent organization and space-time decomposition and distribution in the hosting environment. The behavioral online testing method allows defining the main verification procedures and its preparation, reducing the time and resources costs due to the evolutionary search and multi-agent decomposition of the control analysis, which is important for real-time verification.