区分服务中一种基于聚集流内部公平性的标记算法

A fair aggregate traffic marker for diffserv networks

在现今异构网络Internet,区分服务体系结构通过对不同应用和不同要求的数据流部署不同等级的服务,以提供较大粒度的服务质量(Quality of Service,QoS)保证。不同流之间的公平性问题是区分服务中的研究热点,而标记策略是提高公平性的有效方法。目前许多标记算法只考虑了聚集流之间的公平性,而忽略了聚集流内部流之间的公平性。本文针对同一个聚集流中可能包含不同类型的单个流情况(比如存在自适应TCP流和非适应UDP流、不同速率的多媒体UDP流、采用不同TCP协议的数据流、不同分组大小的数据流),提出了一种基于聚集流内部公平性的标记算法(Fair Aggregate Traffic Marker,FATM)。论文构造了不同情况下的模拟场景,并给出了实验的相关参数设置。大量模拟实验结果表明:标记算法FATM在保持聚集流之间的公平性和网络吞吐量的基础上,提高聚集流内部单个流之间的公平性。

In heterogeneous networks such as today＇s Internet, the differentiated services architecture promises to provide QoS guarantees through scalable service differentiation among different applied and different requirement flows. Fairness issue among different flows is a hot research topic, while marking algorithm is an effective method for improving fairness. Many marking algorithms only consider fairness among the aggregate traffic flows while omitting fairness among the individual flows within an aggregation. Considering real aggregation traffic possibly comprises different type traffics （such as adaptive TCP and non-adaptive UDP traffic, UDP flows with different rates, TCP traffic with different TCP protocol, data traffic with different packet size）, in this paper, we propose a fair aggregate traffic marker （FATM） based on fairness among the individual flows within an aggregation. The paper constructs several simulation scenes considering different condition, and gives the correlative parameter value for the simulation. Simulation results show that FATM can improve the fairness among the individual flows within an aggregation, while keeping the fairness among the aggregate traffic flows and total network throughput.