{"title":"Evaluation of performance and scalability of Mininet in scenarios with large data centers","authors":"J. Ortiz, Jorge Londoño, Francisco Novillo","doi":"10.1109/ETCM.2016.7750830","DOIUrl":null,"url":null,"abstract":"Due to the increased demand for internet services, the challenges that large data centers must handle are bigger. Therefore, in recent years the development of networks defined by software (SDN) has caught the attention of the scientific community, mainly because of the flexibility it presents in its model of centralized management, which facilitates the development of solutions according to demand of the network and / or users. Mininet is one of the most commonly tools used for the study of SDNs, as it allows to recreate data networks that can support various topologies and protocols. Nevertheless, like any other emulation software Mininet has limitations due to the restrictions of the hardware/software platform where it runs, which prevents it to scale to large networks. We present a characterization of the performance and scalability of Mininet via an experimental analysis. For this purpose, we used a typical topology for large data centers. The results let us conclude that, under some restrictions, Mininet gives results in accord with the theoretical expectation. In other cases, the emulation platform exhibits strong dependence on the performance of underlying hardware/software system as well as on the operation mode of the SDN controller used in the network.","PeriodicalId":6480,"journal":{"name":"2016 IEEE Ecuador Technical Chapters Meeting (ETCM)","volume":"60 1","pages":"1-6"},"PeriodicalIF":0.0000,"publicationDate":"2016-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"12","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 IEEE Ecuador Technical Chapters Meeting (ETCM)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ETCM.2016.7750830","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 12
Abstract
Due to the increased demand for internet services, the challenges that large data centers must handle are bigger. Therefore, in recent years the development of networks defined by software (SDN) has caught the attention of the scientific community, mainly because of the flexibility it presents in its model of centralized management, which facilitates the development of solutions according to demand of the network and / or users. Mininet is one of the most commonly tools used for the study of SDNs, as it allows to recreate data networks that can support various topologies and protocols. Nevertheless, like any other emulation software Mininet has limitations due to the restrictions of the hardware/software platform where it runs, which prevents it to scale to large networks. We present a characterization of the performance and scalability of Mininet via an experimental analysis. For this purpose, we used a typical topology for large data centers. The results let us conclude that, under some restrictions, Mininet gives results in accord with the theoretical expectation. In other cases, the emulation platform exhibits strong dependence on the performance of underlying hardware/software system as well as on the operation mode of the SDN controller used in the network.