Transport protocol dependent communications in different packet switch architectures

M. Pustišek, D. Savić, I. Humar, J. Bester
{"title":"Transport protocol dependent communications in different packet switch architectures","authors":"M. Pustišek, D. Savić, I. Humar, J. Bester","doi":"10.1109/MELCON.2006.1653197","DOIUrl":null,"url":null,"abstract":"We investigated the joint influence of different packet switch architectures, controlled and uncontrolled transport layer protocols and transport flow characteristics on the packet switching performance. In network simulator ns2, three packet switch architectures were implemented and combined with UDP and TCP transport. This multilayer model thus combines mechanisms of L1-2 and L4. Additionally, flows with different mean flow size and flow size distribution were applied and the impact transport flow characteristic was analyzed. In particular, the heavy-tail distributed flows were compared to constant flow size. The switching performance measured in terms of packet loss probability and packet delay were analyzed under different traffic loads and buffer sizes. The purpose of the research was thus to understand in more detail a possible impact of new IP traffic patterns (e.g. caused by predominant share of peer-to-peer traffic and corresponding applications) on packet switching systems. The results indicate that the intrinsic characteristics of packet switch architectures reflect in the performance under different transport protocols and flows. The influence of the load or buffer space change is more manageable or predictable, even if we consider the differences among architectures. On the other hand, keeping the same overall load and just replacing the TCP transport with UDP or loading the network with heavy-tail distributed flow sizes or increasing the mean flow size results in more aggressive traffic patterns that may result in a severe deterioration of switching performance","PeriodicalId":299928,"journal":{"name":"MELECON 2006 - 2006 IEEE Mediterranean Electrotechnical Conference","volume":"13 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2006-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"MELECON 2006 - 2006 IEEE Mediterranean Electrotechnical Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/MELCON.2006.1653197","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2

Abstract

We investigated the joint influence of different packet switch architectures, controlled and uncontrolled transport layer protocols and transport flow characteristics on the packet switching performance. In network simulator ns2, three packet switch architectures were implemented and combined with UDP and TCP transport. This multilayer model thus combines mechanisms of L1-2 and L4. Additionally, flows with different mean flow size and flow size distribution were applied and the impact transport flow characteristic was analyzed. In particular, the heavy-tail distributed flows were compared to constant flow size. The switching performance measured in terms of packet loss probability and packet delay were analyzed under different traffic loads and buffer sizes. The purpose of the research was thus to understand in more detail a possible impact of new IP traffic patterns (e.g. caused by predominant share of peer-to-peer traffic and corresponding applications) on packet switching systems. The results indicate that the intrinsic characteristics of packet switch architectures reflect in the performance under different transport protocols and flows. The influence of the load or buffer space change is more manageable or predictable, even if we consider the differences among architectures. On the other hand, keeping the same overall load and just replacing the TCP transport with UDP or loading the network with heavy-tail distributed flow sizes or increasing the mean flow size results in more aggressive traffic patterns that may result in a severe deterioration of switching performance
不同分组交换机体系结构中依赖于传输协议的通信
研究了不同分组交换体系结构、受控和非受控传输层协议以及传输流特性对分组交换性能的共同影响。在网络模拟器ns2中,实现了三种数据包交换体系结构,并将其与UDP和TCP传输相结合。因此,该多层模型结合了L1-2和L4的机制。此外,还采用了不同平均流量和流量分布的流动,分析了冲击输运流动特性。特别地,将重尾分布流与定流尺寸进行了比较。分析了在不同流量负载和缓冲区大小下,以丢包概率和时延衡量的交换性能。因此,研究的目的是更详细地了解新的IP流量模式(例如,由点对点流量的主要份额和相应的应用程序引起的)对分组交换系统的可能影响。结果表明,分组交换体系结构的内在特性反映在不同传输协议和流下的性能上。负载或缓冲区空间变化的影响更易于管理或预测,即使我们考虑了体系结构之间的差异。另一方面,保持相同的总体负载,只是用UDP代替TCP传输,或者用重尾分布流大小或增加平均流大小来加载网络,会导致更激进的流量模式,这可能会导致交换性能的严重恶化
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信