Dynamic Bandwidth Allocation Algorithms In Hybrid Optical-Wireless Broadband Access Networks

Hybrid Optical Wireless Broadband Access Networks (HOWBAN) combining the advantages of optical and wireless access networks is the optimum solution for access networks in future. On one hand, HOWBAN owns high bandwidth and high reliability of optical networks. On the other hand, HOWBAN retains low-cost and flexible of wireless networks. Thus, HOWBAN can offer superior service to users. Some equipments and resources are underused while the network traffic is low, which will waste energy. Therefore, how to design reasonable energy-saving mechanisms is one of the key problems needed to be solved urgently in HOWBAN.At first, the network structure and the research status of HOWBAN are introduced. As well as, the key technologies of HOWBAN are investigated in detail, such as transmission protocol, energy-saving technology, QoS mechanism and so on. Then, this paper analyzes the problem about green Dynamic Bandwidth Allocation (DBA) of Ethernet Passive Optical Network (EPON) at back end. Combined with a sleep control mechanism of Optical Network Units (ONU) and DBA, a green bandwidth allocation algorithm, called Energy-efficient Double-way DBA (EDD), has been proposed. EDD allocates the bandwidth for each ONU according to its upstream and downstream traffic respectively to keep a low delay and to decrease energy consumption. Simulation results show that the proposed EDD can make a trade-off between maximizing the power saving and guaranteeing the network performance.Furthermore, based on the characteristics of HOWBAN and QoS performance of network at front end and back end, a QoS-aware dynamic bandwidth allocation algorithm (QDBA) is also proposed. In QDBA, OLT allocates bandwidth for each queue of ONU-BS according to the traffic and priority of the queue; and each ONU-BS assigns its granted bandwidth to each SS associated with it according to the traffic and priority of each queue in the SS. The simulation results show that, compared with the existing algorithms, QDBA has higher throughput and lower average delay while achieving a better performance of fairness and QoS.