Various algorithms for redistributing tasks in cluster computing systems are described. The results of calculating the probabilistic-time characteristics of the system with connection to the shortest queue and transitions between queues are presented. A number of models with different performance and node failures, with delays in the transition between nodes are described. The results of analytical and simulation modeling of the considered systems are compared.
cluster, load balancing, shortest queue models, queue theory, dispatching, transition between queues, join-the-shortest-queue
1.