Multilevel CLOS data stream exchange algorithm based on smoothed round-robin fair scheduling

Abstract

ABSTRACT The traditional scheduling algorithm has the problems of reducing the success rate of data stream exchange, increasing the exchange time, and increasing the number of required priorities. Therefore, a multi-level CLOS data stream exchange algorithm based on smooth polling fair scheduling is proposed. A multi-level CLOS network data flow scheduling model based on dynamic priority is established, and the value attribute is used as the priority parameter. Design time attribute evaluation criteria to schedule data streams in the best order to ensure that most data streams can be exchanged within the optimal scheduling time. By adjusting the two parameters of data stream value and execution urgency, you can change the degree of influence on the priority of data stream exchange. A smooth cyclic fair scheduling algorithm is used to implement data flow exchange in a multi-stage Clos network system composed of periodic data flows. Experimental results show that the algorithm improves the success rate of data stream exchange, reduces the number of exchanges, reduces the number of exchanges, and reaches the number of priorities required by the system.