与集成服务(InteServ)网络模型不同,区别服务(DiffServ)网络模型中的中心路由器是根据几个有限的服务类别而不是会话流的具体情况来转发调度数据包,这就需要一个重要的服务质量映射的过程。依据DiffServ实现方式的不同,服务质量映射的物理框架也有所不同,但一个关键的映射环节总是存在,即依据数据包的相对优先级系数RPI(RelativePriorityIndex)将其映射入相应的服务类别。该文运用数学工具对这一典型环节进行建模,从视频质量入手分析映射函数,最后在一定条件下得到保证视频质量最优的一种映射函数表达式,并从数学证明和实例分析两个方面论证了这种映射函数表达式的最优性。 Unlike InteServ network models, the central router in the Differentiated Services (DiffServ) network model forwards scheduled packets based on a few limited service classes rather than the specifics of session flow, which requires an important quality of service Mapping process. The physical framework of QoS mapping varies according to DiffServ implementation. However, a key mapping link always exists. That is, the mapping is mapped to the corresponding service class according to the RelativePriorityIndex of the packet. This paper uses mathematical tools to model this typical link, analyzes the mapping function from the video quality, and finally obtains a mapping function expression which ensures the best video quality under certain conditions. From the mathematical proof and case analysis, The paper proves the optimality of this mapping function expression.