Towards opportunistic flow management in OpenFlow

Traditional Internet is merely static, rather than flexible, which is demanded by Internet-Community. OpenFlow has opened the door of management of Internet in a more controlled way. In this context, we are motivated to manage OpenFlow in a more opportunistic way, which encourages flexible network-management by using experience, through trial and error-mechanisms. In this paper, we have proposed a QLearning(Q-L) based opportunistic flow management algorithm OpenFlow-QL, by which NOX(Network Operating System) Controller maintains the flow by considering QoS among OpenFlow switches. We have implemented OpenFlow-QL through a network architecture, where Agent visualizes the opportunistic flow, given by the NOX Controller through VN (Virtual Network) Manager.