Priority data transmission in distributed computing systems

Distributed construction technologies make it possible to scale and increase the fault tolerance of computing systems. However, the stable operation of distributed computing systems (DCS) is hampered by network delays, distributed time management, time spent on synchronization, coordination of system components, and elimination of collisions. Prioritization of data transmission in the DCS makes it possible to ensure a given quality of network service, the characteristics of which include guaranteed throughput, adjustable data bandwidths, and minimal delay for critical data. A model of DCS network using a data transfer scheduler is presented; the influence of data priority, bandwidth, and transmission queues is studied. The model is developed in the AnyLogic environment, which allows to set dynamic and time characteristics (network delays, percentage of data loss, queue balancing). Conclusions are presented on the relationship between the intensity of data entering the network and its priority.