Multi-dimension Qoe Model Based Optimization For Resource Allocation In Hetnets Over SDN

The rapid development of Internet, especially mobile Internet, has brought challenges for mobile operators. To accommodate the heterogeneous networks,the explosive growth of mobile data traffic, the massively connected devices, and a wide variety of novel services,ultra-dense Heterogeneous Networks (HetNets) have been proposed. It is expected that the density of HetNets with overlay coverage through macro cell will increase in the future. To offload the extreme traffic volume in hotspot areas, it is essential to reduce the cell size and add smaller cells. It can also bring the benefits of lower cost, higher efficiency,higher spatial multiplex ratio and the better network performance.For the 5G ultra-dense heterogeneous network, the awareness of environment and user needs along with capabilities of unified resource allocation are required to implement the dynamic network configuration.For example, with the dynamic scheduling among air interfaces transmission technologies in HetNets, the system can make optimal utilization of network resources, thus providing high-quality experiences to the users[1]. Thus, we employ a QoE network model, multi-dimension QoE assessment model, SDN wireless network model and the wireless resource allocation algorithm based on QoE model over SDN.Different from traditional HetNets that mainly deals with interference mitigation and resource allocation between macro cells and small cells, the proposed scheme tries to guarantee the QoE of end users by considering a more comprehensive model of heterogeneity with more flexibility to tradeoff the resource management based on multi-dimension factors impacting QoE, including but not limited to the users'environments and equipment, at the same time, with the capability to enhance the overall potential performance by SDN to adjust the balance of utility among different scales like individual macro cells or small cells,groups of macro cells or small cells.Given that the users subjective evaluation of QoE is affected by multiple factors, we build an extended QoE assessment model to jointly analyze the impact factors from various dimensions, which is called QoE on Demand. These four dimensions includes network conditions, such as network standards; external environment, such as noise level; equipment characteristics, such as screen resolution; user themselves, such as vision.At the same time, these impact factors have some level of independence and correlation. Since the user awareness influences the evaluation of satisfaction directly, to obtain the user awareness factors in advance and improve the quality of network service timely are the key ways to achieve higher communication competitiveness. Therefore, this paper will transform and quantify the multi-dimension factors which affect the quality of communication evaluation to the index of user QoE, which means the QoE demands expression of wireless resource in this case.To face up with HetNets, dense-deployment small cell and differential QoE demands, we propose a dynamic inter-cell wireless resource allocation scheme which expends and provides a more generalized architecture for heterogeneous radio access networks,especially in terms of resource allocation based on QoE requirements with support of SDN. We analyze a typical scenario of imbalanced user demands, the "tidal effect". Analytical and experimental results demonstrate that the proposed scheme has improved user QoE satisfaction due to the high efficiency and extra flexibility of optimization over more dimension of heterogeneity provided by radio resource allocation scheme among SDN in HetNets.