一种基于客户类的接纳控制算法与仿真实现  

Client-class based Admission Control Algorithm and Simulation

作　　者：智英建[1] 王娜[1] 汪斌强[1] 

机构地区：[1]国家数字交换系统工程技术研究中心,郑州450002

出　　处：《系统仿真学报》2008年第24期6686-6690,共5页Journal of System Simulation

基　　金：国家973重点基础研究发展计划(2007CB307102)

摘　　要：接纳控制是视频服务中保证客户服务质量的重要机制。在介绍一种新的分布式视频点播系统的基础上,将客户分为不同的优先级类,每类客户采用基于阈值的接纳控制方法,分析了各个客户类的稳态阻塞概率,并利用Kaufman/Roberts递归算法来降低计算复杂度。建立了系统收益模型,从系统收益最大化的角度,给出了确定各个客户类最佳阈值的方法。仿真结果与理论分析完全吻合,验证了分析模型的有效性和正确性,同时也证明该算法能够保证高优先级的客户优先享用服务器资源,并且就系统收益而言,该算法优于完全共享的接纳策略。Admission control is one important Quality of Service mechanism for maintaining performance of video services. Based on the introduction to a new distributed video-on-demand system, the clients were divided into different priority classes, and the threshold based admission control algorithm was enforced on each class. The blocking probabilities for client classes were analyzed and the Kaufman/Roberts recursive method was used to reduce the calculation complexity. From the view of maximizing system profits, the revenue model was investigated to derive the optimal thresholds for different client classes. The simulation results are fully consistent with the theoretical analyses, which verify the validity and correctness of the proposed model. The results also show that the client with higher priority does have more chances of accessing the server resources, and this scheme performs better than the fully shared admission policy in terms of the system pay-off.

关 键 词：接纳控制 服务质量 视频点播 阻塞概率 系统收益 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]