Study On The Key Technologies Of The Convergence Of Pon And Mobile Access Network

With the development of various broadband access technologies and ever-increasing diversity of communication needs from users, access network shows a development trend of high bandwidth and convergence of kinds of acess technologies. On one hand, the mobile access network is characterized by wide coverage, flexible mobitily, but is faced with some challenges such as cost reduction and bandwidth increasing. On the other hand, Passive Optical Network (PON) is characterized by its high bandwidth, full-service access, high efficient QoS mechanism, lower cost, but it is limited by its lack of mobility. Therefore, the convergence of PON and mobile access network is gaining more and more attentions.This thesis mainly studies on the key technologies of the convergence of PON and mobile access network including resources allocation in the converged network, bandwidth allocation in LTE backhaul network based on XGPON supporting direct communications between eNodeBs, evaluation of user experience for mobile video services, joint bandwidth allocation in Wi-Fi access network and resources allocation in hybrid network of3G and Wi-Fi network, etc, and gains a series of innovative research achievements. The main work and innovative contribution are lists as follows.1. In the converged network of PON and mobile access network, the independent allocation of bandwidth in PON and wireless spectrum brings low throughput and large delay. To solve these problems, a joint resources allocation algorithm in the converged network of PON and mobile access network is proposed, which dynamically adjusts the guaranteed bandwidth of converged ONU (ONU with function of BS) according to the change of spectrum allocation. The simulation results show that the proposed algorithm can improve throughput by10%and decreased the end-to-end delay by one time.2. To meet the requirement of low-delay communications between eNodeBs in LTE backhaul network, a dyanamic bandwidth allocation (DBA) algorithm supporting low-delay communications between eNodeBs in LTE backhaul network based on XGPON is proposed, which conducts two-steps bandwidth allocation in converged OLT (OLT with function of MME/S:GW) and converged ONU, respectively. The simulation results indicate that when the ratio of services from eNodeB to MME/S-GW to inter-eNodeB services is8:2, comparing with traditional backhaul network which requires the inter-eNodeB services to be transfered in MME/S-GW, the proposed algorithm improves throughput by12%in heavy network load; the average delay of communications between eNodeBs under proposed algorithm is decreased by several times; the proposed algorithm provides better fairness in communication links between eNodeBs.3. To solve the problems that the performance evaluation index for traditional video services cannot exactly reflect the user experience for mobile video services, a performance metric for mobile video services based on user experience has been proposed-User Experience Index (UEI). And the UEI reflects user experience of video services by considering three aspects of video play time, video play fluency and user join time.4. In order to improve the user experience for mobile video services in Wi-Fi access network with multiple APs, a joint APs bandwidth allocation algorithm based on greedy method is proposed. Furthermore, a cost performance algorithm is proposed based on the above algorithm. The simulation results show that greedy-based joint bandwidth allocation algorithm improves UEI by two times than the independent AP bandwidth allocation algorithm, and the cost performance algorithm achieves better UEI than greedy algorithm by50%.5. To improve user experience of mobile video services in such hybrid network that Wi-Fi network with multiple APs and3G network, a resources of AP bandwidth and3G cellular usage allocation algorithm based on Divide and Conquer Strategy (DCS) is proposed. And an improved algorithm is further proposed based on the above algorithm, which uses Integrated and Cooperate Strategy (ICS). The simulation results show that comparing with random algorithm, DCS improves the performance by30%, and the ICS improves better performance than DCS by30%. Furthermore, based on the above research, a prediction-based ICS algorithm is proposed to improve the ICS algorithm by considering the scenario of dynamic user arrival. The prediction-based ICS algorithm is to dynamically allocate bandwidth and3G celluar usage based on the historical data and request of current user. The simulation results show that the proposed algorithm improves user experience of mobile video services by38%in the hybrid network of Wi-Fi and3G network.