DISTRIBUTED COMPUTER SIMULATION IN CLOUD COMPUTING SYSTEMS

Abstract

The article examines issues of increasing the efficiency of distributed modeling systems in a cloud virtual environment. One of the main tasks that appears in the process of modeling complex systems is the scaling of calculations – increasing the number of resources when the size of the task increases. The paper examines the models used for computer simulation, in particular the serial and parallel model with different synchronization methods. A cloud-based model of the simulation environment with the use of thin clients, decentralized control and asynchronous synchronization is considered, which allows scaling the simulation system in the horizontal direction. This model allows flexible management of resources depending on the complexity of the modeling object, dynamic assignment and release of computing virtual nodes. Experiments on modeling 3D objects on different number of virtual machines are described in the paper. The use of many nodes, on the one hand, increases the speed of calculations, on the other hand, it requires a lot of computing power for synchronization, management and exchange of messages between nodes, so the dependence of the increase in speed when using additional cloud nodes is not linear. The involvement of a distributed simulation environment made it possible to double the size of the simulation in the case of using 4 computing nodes. An increase in productivity was observed with an increase in the number of computing nodes. The results of the research can be used in the development of new methods of resource allocation and technologies of distributed computing in cloud systems, models of efficient management of computing nodes, systems of distributed simulation modeling.