Dynamic Load Balancing Model with Time Delay in Distributed Environment

Load distribution of distributed system is not balanced. The imbalance could cause low efficiency. There are many studies on how to balance loads in a distributed system, among which are multiple studies achieve the goal by using linear time delay system. In the paper, we propose an impulsive and switching load balancing model with time delay based on control theory. The model aims at dynamics of resource and task running status. In order to describe various current states of a node, we construct corresponding subsystem according to the dynamics of node resources. The subsystem switching is triggered by impulsive signal which can decrease the communication overhead among nodes. The model reallocates tasks in light of task's real-time running status so as to improve the efficiency of dynamic load balancing. Experimental results demonstrate that the proposed model can make the system balanced, and verify the feasibility of the model.