PRACTICE-ORIENTED TECHNOLOGY FOR MODELING LARGE-SCALE ORGANIZATIONAL SYSTEMS (1000+ PARTICIPANTS) USING 1C LOAD TESTING TOOLS

The article describes main tasks, the solution of which is required to bring the system of multiagent simulation modeling of organizational behavior, using load testing approaches on the 1C platform, to the level of real requirements of the subject area (more than 1000 participants). It is required to solve the issues of the next level for a system that correctly models process with a small number of participants. The technology described a solution to four of them, listed below. Using server background jobs as carriers of model agents has its advantages over client applications. These are lower hardware requirements and availability of free licenses. The scaling the number of participants includes options with and without the use of participant profiles. The choice of the most appropriate one is justified based on minimal model correction needs, acceptable model behavior predictability and enough model versatility.Besides using time in the main model queues and tasks, participants forced to spend time in queues in the carrier system (it calls technological noise). The more participants there are, the more noise there is. The variant of reducing the technological noise is partial refusal of using transactions and transition from transactional locks upon areas to non-transactional locks upon objects. It both reduces transactions durations and increases lock granularity. The high number of participants and some freedom in task accepting mechanisms produces race conditions. It is possible to fight of undesirable ones by limitation to take additional tasks for executors and by limitation to assign work to profiles rather than to their concrete representatives. Summary: the described technology allows obtaining data for models with the required number of participants. It meets the objectives of the stated subject area (development, implementation and maintenance of IT solutions) and has possibilities for expansion beyond it.