一种新型的基于预测的卫星资源分配方法

针对传统的按需分配方式所造成的长时延问题,提出了一种新型的基于业务预测的按需分配方法。该方法通过预测业务在等待队列中的长度,将预测信息发送给网络资源控制中心,再由控制中心动态地分配带宽资源,达到服务质量的要求。在阐述了按需分配和业务预测的基本算法后,通过仿真分析了不同的业务类型采用不同的预测信息发送间隔时间的时延性能,分析中考虑了实时业务比非实时业务享有更高的优先权,并且比较了常规按需分配与基于业务预测的按需系统分配方法的时延性能。仿真结果表明利用业务预测,选择合适的预测信息发送间隔时间,能够有效地改善带宽资源分配的时延性能,相比一般的按需分配方式服务质量有了明显的提高。

Novel Resource Allocation Scheme Based on Traffic Prediction in Satellite Systems

A novel dynamic bandwidth management mechanism based on the traffic prediction is presented to solve the problem of long delay caused by traditional demand assigned multiple access(DAMA).Firstly,the satellite terminals predict the queue length of the buffer,then send the reports to the network control center(NCC).Finally,the resources are allocated to the terminals by NCC.The algorithm performance is validated through the simulation for different types of traffic and different intervals between two consecutive prediction reports.The priority of real time traffic is considered.The delays of DAMA based on prediction and normal case are compared.The theoretical analysis and simulation results show that the new DAMA algorithm can efficiently tackle the problem of delay between bandwidth request and bandwidth assignment.