Soft Aggregation of Multimedia Flows Based on QoS Classes

Abstract

In the software defined network (SDN) /OpenFlow architecture, the existing flow aggregation methods depend heavily on quantitative data and thresholds. That is, the flows should have explicit values and weights for all Quality of Service (QoS) parameters, and the QoS classes in the system should be fixed, etc. However, in reality, the above information is typically imprecise and time-varying, especially in wireless networks. In this paper, based on preference logic, an SDN network traffic aggregation method, termed SAM (soft aggregation model), is proposed to model the imprecise QoS requirements of multimedia flows, which achieves a soft flow aggregation without strict thresholds. The experiment results show that the soft method of SAM can effectively adjust the aggregation process to make best use of system resources. Therefore, it performs better than the existing aggregation methods in terms of throughput.