无线Internet的业务质量及其改进

在移动网络中,移动用户对Internet的访问,具有与通常的外部Internet网络不同的特点。移动用户发出的对信息内容的请求(例如询问前方交通情况和路线),以及无线Internet信息提供商主动向移动用户发出的信息(例如移动用户附近的商品信息广告),与移动用户的位置、移动状态(方向、速度)和个人情况(userprofile)等有关。因此,要建立恰当的无线Internet流量模型和实现QoS(业务质量)控制,需要考虑用户的移动特性和在不同移动状态下访问Internet的概率分布。该文将建立一个新的定量分析模型,用半马尔可夫过程和一般的状态空间来描述移动用户的行为;并基于该移动模型,建立相应的移动用户对信息内容的访问模型、描述移动用户请求的到达过程;进而得到无线Internet业务质量参数的估计,即无线Internet对移动用户请求的平均响应时间、对无线侧和有线侧传输带宽的需求等;利用这些结果,可以实现无线资源的动态分配和允入控制。该文还将分析无线代理网关的性能、缓存容量需求,并给出改善平均响应时间的可行方案。最后用计算机模拟对提出的理论分析进行验证。

Quality of Wireless Internet Services and its Improvement

In a wireless network,the movement of mobile users presents a scenario of access to the Internet that is substantially different from the wired network.Content requested by a mobile user or pushed by a wireless Internet ser-vice provider depends on the mobile user's location,moving state(e.g.,velocity,direction)and its profile.Therefore,we need to consider the probability with which mobile users access Web pages.In this paper,we first construct a new ana-lytical/numerical model that characterizes mobile user behavior in a general state-space using a semi-Markov process representation.Based on the mobility model and the resultant request model,we analyze the content access patterns in a wireless Internet and then obtain estimates of its quality of services,including the total average response time ,and bandwidth resources required for the wireless and wired network.We then evaluate the performance of a wireless proxy gateway,analyze its capacity requirement ,and formulate the wireless resource allocation.A prefetch scheme is proposed to improve the quality of wireless Internet services.Finally,some numerical results are presented.