{"title":"Controller placement strategies for a resilient SDN control plane","authors":"P. Vizarreta, C. M. Machuca, W. Kellerer","doi":"10.1109/RNDM.2016.7608295","DOIUrl":null,"url":null,"abstract":"Software Defined Networking (SDN) provides efficient network control and management by introducing a logically centralized control plane. However, outsourcing of the control plane intelligence to an SDN controller requires reliable switch-to-controller connection. This paper presents two strategies to address the Reliable Controller Placement (RCP) problem, which protect the control plane against single link and node failures and provide seamless failover by exploiting the principles of resilient routing. The first approach considers that switches have to be connected to a controller over two Disjoint Control Paths (RCP-DCP). The second approach considers that switches have to be connected to two Different Controller Replicas (RCP-DCR) over two disjoint paths. Both approaches are finding working and protection control paths of minimum length to enable fast and efficient failover. The two models have been compared with respect to the unprotected scenario, in terms of control path length, expected control path loss in different failure scenarios and average control path availability. The results show that both models RCP-DCP and RCP-DCR improve significantly the resilience of the control plane, while adding very limited penalty to the average control path length. If link failures are the dominating failures, both RCP-DCP and RCP-DCR offer similar performance and the best strategy depends on the topological characteristics and the number of the controllers in the network. RCP-DCR provides additional protection against controller failures and it shows better performance when node failures are dominating or are comparable to link failures.","PeriodicalId":422165,"journal":{"name":"2016 8th International Workshop on Resilient Networks Design and Modeling (RNDM)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"76","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 8th International Workshop on Resilient Networks Design and Modeling (RNDM)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/RNDM.2016.7608295","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 76
Abstract
Software Defined Networking (SDN) provides efficient network control and management by introducing a logically centralized control plane. However, outsourcing of the control plane intelligence to an SDN controller requires reliable switch-to-controller connection. This paper presents two strategies to address the Reliable Controller Placement (RCP) problem, which protect the control plane against single link and node failures and provide seamless failover by exploiting the principles of resilient routing. The first approach considers that switches have to be connected to a controller over two Disjoint Control Paths (RCP-DCP). The second approach considers that switches have to be connected to two Different Controller Replicas (RCP-DCR) over two disjoint paths. Both approaches are finding working and protection control paths of minimum length to enable fast and efficient failover. The two models have been compared with respect to the unprotected scenario, in terms of control path length, expected control path loss in different failure scenarios and average control path availability. The results show that both models RCP-DCP and RCP-DCR improve significantly the resilience of the control plane, while adding very limited penalty to the average control path length. If link failures are the dominating failures, both RCP-DCP and RCP-DCR offer similar performance and the best strategy depends on the topological characteristics and the number of the controllers in the network. RCP-DCR provides additional protection against controller failures and it shows better performance when node failures are dominating or are comparable to link failures.
SDN (Software Defined Networking)通过引入逻辑上集中的控制平面,提供高效的网络控制和管理。但是,将控制平面智能外包给SDN控制器需要可靠的交换机到控制器连接。本文提出了两种解决可靠控制器放置(RCP)问题的策略,它们利用弹性路由原理保护控制平面免受单链路和节点故障的影响,并提供无缝故障转移。第一种方法认为交换机必须通过两条分离控制路径(RCP-DCP)连接到控制器。第二种方法认为交换机必须通过两条不相交的路径连接到两个不同的控制器副本(RCP-DCR)。这两种方法都在寻找最小长度的工作和保护控制路径,以实现快速有效的故障转移。在控制路径长度、不同故障情况下的预期控制路径损失和平均控制路径可用性方面,将两种模型与未受保护的情况进行了比较。结果表明,RCP-DCP和RCP-DCR模型均显著提高了控制平面的弹性,同时对平均控制路径长度的影响很小。如果链路故障是主要故障,RCP-DCP和RCP-DCR都提供类似的性能,最佳策略取决于拓扑特征和网络中控制器的数量。RCP-DCR提供了针对控制器故障的额外保护,当节点故障占主导地位或与链路故障相当时,它显示出更好的性能。