qdfsn: qos-enabled dynamic and programmable framework for sdn

Software defined network (sdn) can integrate a lot of network functions such as network resource management into a consolidated framework. tcp operates in these networks with low information traffic characteristics. as a result, it has to continuously change its congestion window size in order to handle drastic changes in the network or its traffic conditions. as a result, tcp frequently overshoots or undershoots its transmission rate, making it a native congestion control protocol. to overcome that problem, we have proposed a new qos framework for sdn called qdfsn (qosenabled dynamic and programmable framework for sdn) which can be effectively applied in data centers as well. in this, and by means of aqm (active queue management), a new function for detecting the upcoming congestion situation is designed. in each node, a developed mathematical model is used to calculate the best parameters of the node adaptively, especially the service rate, to minimize the congestion in the network. this model is tested in many ns2 scenarios, and the results are presented. the results show improvements in selected qos parameters like throughput and delay. we conclude that qdfsnbased congestion control shortens the process of adapting tcp to network circumstances, and enhances the tcp performance.

QDFSN: QoS-enabled Dynamic and Programmable Framework for SDN

Software Defined Network (SDN) can integrate a lot of network functions such as network resource management into a consolidated framework. TCP operates in these networks with low information traffic characteristics. As a result, it has to continuously change its congestion window size in order to handle drastic changes in the network or its traffic conditions. As a result, TCP frequently overshoots or undershoots its transmission rate, making it a native congestion control protocol. To overcome that problem, we have proposed a new QoS framework for SDN called QDFSN (QoSenabled Dynamic and Programmable Framework for SDN) which can be effectively applied in Data Centers as well. In this, and by means of AQM (Active Queue Management), a new function for detecting the upcoming congestion situation is designed. In each node, a developed mathematical model is used to calculate the best parameters of the node adaptively, especially the service rate, to minimize the congestion in the network. This model is tested in many NS2 scenarios, and the results are presented. The results show improvements in selected QoS parameters like throughput and delay. We conclude that QDFSNbased congestion control shortens the process of adapting TCP to network circumstances, and enhances the TCP performance.