{"title":"Pathseer: a centralized tracer of packet trajectories in software-defined datacenter networks","authors":"Amer Aljaedi, C. E. Chow","doi":"10.1109/IPTComm39427.2016.7780246","DOIUrl":null,"url":null,"abstract":"Tracing packet trajectories is a key component of network fault diagnosis, especially in software-defined datacenter networks. The network operators and SDN developers often need to reason about the network forwarding behavior and localize the unpredicted network failures. The existing approaches for tracing packet trajectories in an SDN environment either incur large data collection overhead or trade off the data plane resources to satisfy pre-specified queries. This paper introduces PathSeer, a lightweight and scalable technique for tracing packet trajectories in SDN-enabled datacenters. PathSeer leverages OpenFlow features to reuse some fields in the packet header for encoding the path of the traced packet. It is designed as an SDN tool running on the centralized network controller. Therefore, PathSeer can efficiently trace the end-to-end paths taken by packets, and provide insights into the distributed forwarding state of the network. A preliminary evaluation on a fat-tree topology shows that PathSeer significantly reduces the overhead of tracing packet trajectories, compared to state-of-the-art techniques.","PeriodicalId":302222,"journal":{"name":"2016 Principles, Systems and Applications of IP Telecommunications (IPTComm)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 Principles, Systems and Applications of IP Telecommunications (IPTComm)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IPTComm39427.2016.7780246","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 7
Abstract
Tracing packet trajectories is a key component of network fault diagnosis, especially in software-defined datacenter networks. The network operators and SDN developers often need to reason about the network forwarding behavior and localize the unpredicted network failures. The existing approaches for tracing packet trajectories in an SDN environment either incur large data collection overhead or trade off the data plane resources to satisfy pre-specified queries. This paper introduces PathSeer, a lightweight and scalable technique for tracing packet trajectories in SDN-enabled datacenters. PathSeer leverages OpenFlow features to reuse some fields in the packet header for encoding the path of the traced packet. It is designed as an SDN tool running on the centralized network controller. Therefore, PathSeer can efficiently trace the end-to-end paths taken by packets, and provide insights into the distributed forwarding state of the network. A preliminary evaluation on a fat-tree topology shows that PathSeer significantly reduces the overhead of tracing packet trajectories, compared to state-of-the-art techniques.