QoS enhancement using embedded agent in OpenfLow based SDN node

In Software Defined Network (SDN) flow rules for data packets in a session are decided in controller-nodes which are geographically away from data carrying nodes. The traverse time for request and control packets between controller-node and data-node causes delay in executing flow rule on packets. Flow rules are stored in data node for limited time to save expensive memory. Whenever some flow rule expires, which is still required for packet forwarding, controller needs to be consulted again by data node. Frequency of above consultation affects QoS. We propose a novel concept by introducing embedded agent in data-node and running our algorithm on it to reduce the frequency of consultation to controller by data-node on expiry of flow rule, which in turn improves QoS. We have implemented proposed algorithm in Network Lab. of our University and results of simulator and test bed reveal improvement of QoS. We have also derived an analytical model to find delay between controller and data node. Test results shows that there is 66.6% saving in controller load and sufficient saving in packet loss depending on network.