CAREER: Adaptive Admission Control for Broadband Communication

这项研究将制定和开发一种用于控制宽带通信网络上承载的业务的方法，特别是对于承载各种实际业务源的情况，这些业务源具有难以建模和分析的排队属性；源之间的相关性；缓慢变化的统计属性；多样化的异构源类型；错误指定的业务描述符；以及疏忽的网络业务整形。重点是当业务源被给予服务质量保证并且目标是最大化利用网络资源的情况。该方法基于统计分类技术，该技术使用历史数据来确定哪些是可接受的承载流量组合，哪些不是。解决的问题包括历史数据的有效表示和存储；分类器模型的选择和修改策略；数据上的噪声；决策的置信度区间；决策空间的定向和非定向探索；过滤掉无信息的数据；以及实现。开发的方法的有效性和优势将在简单模型和真实流量轨迹(如以太网数据、可变速率视频和ISDN流量)上与替代方法进行实验测试。

This research will formulate and develop a method for controlling the traffic carried on a broadband communication network, specifically for the case of carrying diverse real traffic sources with difficult to model and analyze queueing properties; inter-source correlation; slowly varying statistical properties; diverse heterogeneous source types; mis-specified traffic descriptors; and inadvertent network traffic shaping. The focus is the case when the traffic sources are given quality of service guarantees and the goal is to maximize utilization of network resources. The method is based on statistical classification techniques that use historical data as to what were acceptable combinations of carried traffic and what were not. Problems addressed include efficient representation and storage of historical data; choice and modification strategy of classifier model; noise on the data; confidence intervals on decisions; directed and undirected exploration of the decision space; filtering out uninformative data; and implementations. The effectiveness and advantages of the methods developed will be experimentally tested against alternative methods on simple models and real traffic traces such as ethernet data; variable rate video; and ISDN traffic.