Techniques Study On Resource Control At The Edge Of The Access Network In NGB Open Service Platform

In recent years, China has begun to promote triple nework convergence of the telecommunications network, Internet and cable television network. The ongoing triple nework convergence strategy allows operators to operate services which were restricted by government regulations in the past and enable them to become full service/media operators. As the two-way transformation of cable television network and the implementation of next generation broadcast networks strategy continues, it is urgent for cable operators to build a unified service platform to open the service abilities of the edge of the access network to support the delivery of services after triple nework convergence and to ensure that services can be managed and controlled.Under the support of the National Science and Technology Support Programnamed"next generation radio and television service system that support cross-domainand multi-operators",this dissertation mainly studies the technology of 3TNet Open Platform. In particular, we focus on the related technologies about the resource control at the edge of the access network.The main contributions of this dissertation can be summarized as follows:1. A new resource management method based on revenue mechanism is proposed. First, we analyze the behaviors of four typical services in television networks and obtain their revenue functions. For non-cross-domain environment, we use a greedy algorithm utilizing marginal revenues to allocate resources at the edge of the access network in order to maximize the revenue of operators while satisfying the minimum requirements of QoS. For cross-domain environment, the two open service platforms, which collaborate to provide users with services, negotiate their revenue functions and allocate their own resources with the same method as the resource planning method for non-cross-domain environment.2. A new admission control framework at the edge of the access network is proposed and a novel QAM resource allocation algorithm for VOD service under the new framework is given. The new framework uses two types of strategies: the admission decision strategies which aim to decide whether to accept the service request and the frequency selection strategies which aim to decide which frequency should be allotted to the service request. This framework is flexible as it allows adjusting of old admission control strategies and adding of new admission control strategies. The novel algorithm groups all frequencies in accordance with the size of remaining bandwidth and selects a frequency in accordance with the order of first select a group and then selects the frequency when service request arrives. As the algorithm ensures the support capacity to the follow-up service requests as much as possible, it achieves a lower service blocking probability than other QAM resource allocation algorithms, and can guarantee a certain degree of load balance.3. A novel workflow scheduling algorithm based on decomposition of QoS is proposed. By dividing the global QoS constraint of the whole workflow into local QoS constraints of individual nodes, the new algorithm reduces the complexity of the problem. This new algorithm is more efficient and can find a better solution compared to other algorithms based on decomposition of global QoS constraint.