摘要
针对目前在流量整形器建模中将整形器视为无限缓存设备的缺陷,基于网络演算,使用最小加代数建立了有限缓冲区的流量整形器(FSS)模型,获得了FSS的分组时延和分组丢失与预留缓存空间的关系,给出了FSS性能参数的最小加代数表示.研究结果表明:当贪心整形器的服务曲线大于业务流的到达曲线时,整形器的引入不会额外增加业务流丢失的分组数,而整形器的缓冲特性能够减少网络中业务流丢失的分组数;在给定目标服务质量参数的前提下,相关结论可用于确定资源预留的上界,以改进网络的规划与设计.
Aiming at the flaw that the shaper is regarded as an unlimited storage device when modelling the flow shaper, a finite storage shaper （FSS） model was built by min-plus algebra based on network calculus. The relation between the packet delay and packet loss of FSS and the reserved buffer space was obtained and the FSS performance parameters were represented by min-plus algebra. Research results show that the shaper will not extra increase the packet number of flow loss when the greedy service curve of shaper is larger than the arrival curve of the flow, and the buffering characteristic of the shaper can decrease the packet number of the flow loss in the network. The results can be applied to determining the upper bound of resource reservation so as to improve the design of network when the target parameters of quality of service are given.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2005年第10期1068-1071,共4页
Journal of Xi'an Jiaotong University
基金
国家高技术研究发展计划资助项目(863-511-946-008)
作者简介
高岭（1972-），男，博士生；
李增智（联系人），男，教授，博士生导师．
