On the performance analysis of traffic splitting on load imbalancing and packet reordering of bursty traffic

S. Prabhavat, Hiroki Nishiyama, N. Ansari, N. Kato
{"title":"On the performance analysis of traffic splitting on load imbalancing and packet reordering of bursty traffic","authors":"S. Prabhavat, Hiroki Nishiyama, N. Ansari, N. Kato","doi":"10.1109/ICNIDC.2009.5360896","DOIUrl":null,"url":null,"abstract":"Owing to the heterogeneity and high degree of connectivity of various networks, there likely exist multiple available paths between a source and a destination. To be able to simultaneously and efficiently use such parallel paths, it is essential to facilitate high quality network services at high speeds. So, traffic splitting, having a significant impact on quality of services (QoS), is an important means to achieve load balancing. In general, most existing models can be classified into flow-based or packet-based models. Unfortunately, both classes exhibit some drawbacks, such as low efficiency under the high variance of flow size in flow-based models and the phenomenon of packet reordering in packet-based models. In contrast, Table-based Hashing with Reassignment (THR) and Flowlet Aware Routing Engine (FLARE), both belonging to the class of flow-based models, attempt to achieve both efficient bandwidth utilization and packet order preservation. An original flow can be split into several paths. As compared to the traditional flow-based models, load balancing deviation from ideal distribution decreases while the risk of packet reordering increases. In this paper, we introduce analytical models of THR and FLARE, and derive the probabilities of traffic splitting and packet reordering for each model. Our analysis shows that FLARE is superior to THR in packet order preservation. Also, the performance of FLARE on bursty traffic is demonstrated and discussed.","PeriodicalId":127306,"journal":{"name":"2009 IEEE International Conference on Network Infrastructure and Digital Content","volume":"4 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2009-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"10","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2009 IEEE International Conference on Network Infrastructure and Digital Content","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICNIDC.2009.5360896","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 10

Abstract

Owing to the heterogeneity and high degree of connectivity of various networks, there likely exist multiple available paths between a source and a destination. To be able to simultaneously and efficiently use such parallel paths, it is essential to facilitate high quality network services at high speeds. So, traffic splitting, having a significant impact on quality of services (QoS), is an important means to achieve load balancing. In general, most existing models can be classified into flow-based or packet-based models. Unfortunately, both classes exhibit some drawbacks, such as low efficiency under the high variance of flow size in flow-based models and the phenomenon of packet reordering in packet-based models. In contrast, Table-based Hashing with Reassignment (THR) and Flowlet Aware Routing Engine (FLARE), both belonging to the class of flow-based models, attempt to achieve both efficient bandwidth utilization and packet order preservation. An original flow can be split into several paths. As compared to the traditional flow-based models, load balancing deviation from ideal distribution decreases while the risk of packet reordering increases. In this paper, we introduce analytical models of THR and FLARE, and derive the probabilities of traffic splitting and packet reordering for each model. Our analysis shows that FLARE is superior to THR in packet order preservation. Also, the performance of FLARE on bursty traffic is demonstrated and discussed.
突发流量中负载不均衡和数据包重排序的流量分割性能分析
由于各种网络的异构性和高度连通性,在源和目的之间可能存在多条可用路径。为了能够同时有效地利用这些并行路径,必须提供高质量的高速网络服务。因此,对服务质量(QoS)有重大影响的流量分割是实现负载均衡的重要手段。一般来说,大多数现有的模型可以分为基于流的模型和基于包的模型。遗憾的是,这两种方法都存在一些缺点,例如基于流的模型在流量大小变化较大的情况下效率较低,以及基于包的模型中存在数据包重排序现象。相比之下,基于表的重分配哈希(THR)和流量感知路由引擎(FLARE),都属于基于流的模型,试图实现有效的带宽利用和包顺序保存。一个原始流可以被分割成几个路径。与传统的基于流的模型相比,该模型减少了负载均衡与理想分布的偏差,但增加了数据包重排序的风险。本文介绍了THR和FLARE的分析模型,并推导了每个模型的流量分裂和数据包重排序的概率。分析表明,FLARE在包序保存方面优于THR。同时,对FLARE在突发业务中的性能进行了演示和讨论。
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