Adaptive Traffic Offloading Method Based On SDN In Heterogeneous Networks

With the rapid development of communication technology, wireless data is increasing greatly, and traffic overload problem in the cellular networks has become crucial. In addition, considering the integration of heterogeneous networks is an inevitable trend in the future wireless communication, and various access networks have different bandwidth, frequency, specification and quality of service. If the system lacks effective management, it may lead the uneven distribution of network resources and lowly spectral efficiency. Therefore, to make sure that the cooperation between heterogeneous networks in traffic area is good, it is necessary to study more effective load-balancing ways.Focusing to traffic offloading in heterogeneous network, we have studied the new network architecture and the way to realize adaptive offloading. The main points are as follows:First of all, based on the concept of software defined network (SDN) and integration scenarios of different networks, this paper has proposed a three-layer structure, and designed the controller and south protocol. Besides, the resources in wireless network are aggregated by a specific method of abstracting heterogeneous wireless network. Providing an open API between the control plane and the application plane, the heterogeneous network can be managed consistently and efficiently.Then, an adaptive traffic offloading method based on a biologically inspired attractor selection model is proposed. The method considers a queuing model based on discrete Markov modulated Poisson process (dMMPP) to obtain the real-time network status in LTE and WLAN, respectively. Then, applying the attractor selection model to the application layer through open interface, which is between control plane and data plane, the traffic offloading is performed under the management of the controller. At last, the simulation results demonstrate that the proposed mechanism can realize the offloading between LTE and WLAN. Besides, it can also balance the system delay and the handover times, in order to satisfy the traffic request of users and improve the system stability.