Abstract
The availability of high-performance computing, high speed data transfer over the network and widespread of software for the design and pre-production in mechanical engineering have led to the fact that at the present time the large industrial enterprises and small engineering companies implement complex computer systems for efficient solutions of production and management tasks. Such computer systems are generally built on the basis of distributed heterogeneous computer systems. The analytical problems solved by such systems are the key models of research, but the system-wide problems of efficient distribution (balancing) of the computational load and accommodation input, intermediate and output databases are no less important. The main tasks of this balancing system are load and condition monitoring of compute nodes, and the selection of a node for transition of the user’s request in accordance with a predetermined algorithm. The load balancing is one of the most used methods of increasing productivity of distributed computing systems through the optimal allocation of tasks between the computer system nodes. Therefore, the development of methods and algorithms for computing optimal scheduling in a distributed system, dynamically changing its infrastructure, is an important task.
Highlights
Currently, the design and production preparation in mechanical engineering implemented a wide range of engineering calculations
Static balancing is performed prior to the start of the task, semi-dynamic load balancing requires that the distribution of tasks carried out at the stage of initialization and dynamic balancing is performed in the course of computation
If we talk about how to interact the load balancing in a computer system, the balancing is divided into central, where the special compute node distributes the modules of a task between calculators, and distributed, in which status data is exchanged between all nodes in a distributed computing system by using a special algorithm [1,2,3,4,5,6,7,8,9,10,11,12,13,14]
Summary
The design and production preparation in mechanical engineering implemented a wide range of engineering calculations (strength, hydrodynamic, thermal, etc.) These tasks require a much more machine resources than the systems of design or production preparation. Static balancing is performed prior to the start of the task, semi-dynamic load balancing requires that the distribution of tasks carried out at the stage of initialization and dynamic balancing is performed in the course of computation. The article describes the method of computational load balancing in distributed automation systems which are focused on multi-agent and multi-threaded data processing. A GPSS-model and the results of model experiments for developed algorithm load dispatching are represented This model and results show the effectiveness of the algorithm even with a significant increase in the number of connected nodes and with an increase in architecture of distributed computing system.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
