{"title":"特定于目标的最大冗余树:设计、性能比较和应用程序","authors":"Wolfgang Braun, Daniel Merling, M. Menth","doi":"10.1109/ICIN.2018.8401580","DOIUrl":null,"url":null,"abstract":"Resilient networks react to a failure by reconfiguring routing tables in a network-wide manner and utilizing fast reroute (FRR) in the meanwhile. FRR uses pre-computed backup paths to avoid the failure locally. The existing FRR approaches for IP networks and Software-Defined Networking (SDN) differ in regard to computational complexity, backup path length, additional forwarding state and required capacities. Maximally Redundant Trees (MRTs) are standardized by the IETF and are already available for IP and MPLS networks. In this paper, we propose destination-specific MRTs (dMRTs), an improved variant of MRTs, for protection of IP networks. This mechanism reduces the generated path lengths but increases the computational complexity of the MRT algorithm. We analyze and compare our proposal with simple MRTs and MPLS FRR. We show on a large and publicly available network data base that dMRTs provide significant shorter backup paths compared to simple MRTs and require less forwarding state and network capacities than MPLS FRR. Adapting an existing FRR mechanism for IP networks to SDN enables its deployment in hybrid-SDN networks. Thus, we outline implementation options of dMRTs in IP and MPLS networks and (d)MRTs in SDN. Finally, we discuss the computational overhead of our proposal in the context of (decentralized) IP networks and (centralized) SDNs.","PeriodicalId":103076,"journal":{"name":"2018 21st Conference on Innovation in Clouds, Internet and Networks and Workshops (ICIN)","volume":"187 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2018-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Destination-specific maximally redundant trees: Design, performance comparison, and applications\",\"authors\":\"Wolfgang Braun, Daniel Merling, M. Menth\",\"doi\":\"10.1109/ICIN.2018.8401580\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Resilient networks react to a failure by reconfiguring routing tables in a network-wide manner and utilizing fast reroute (FRR) in the meanwhile. FRR uses pre-computed backup paths to avoid the failure locally. The existing FRR approaches for IP networks and Software-Defined Networking (SDN) differ in regard to computational complexity, backup path length, additional forwarding state and required capacities. Maximally Redundant Trees (MRTs) are standardized by the IETF and are already available for IP and MPLS networks. In this paper, we propose destination-specific MRTs (dMRTs), an improved variant of MRTs, for protection of IP networks. This mechanism reduces the generated path lengths but increases the computational complexity of the MRT algorithm. We analyze and compare our proposal with simple MRTs and MPLS FRR. We show on a large and publicly available network data base that dMRTs provide significant shorter backup paths compared to simple MRTs and require less forwarding state and network capacities than MPLS FRR. Adapting an existing FRR mechanism for IP networks to SDN enables its deployment in hybrid-SDN networks. Thus, we outline implementation options of dMRTs in IP and MPLS networks and (d)MRTs in SDN. Finally, we discuss the computational overhead of our proposal in the context of (decentralized) IP networks and (centralized) SDNs.\",\"PeriodicalId\":103076,\"journal\":{\"name\":\"2018 21st Conference on Innovation in Clouds, Internet and Networks and Workshops (ICIN)\",\"volume\":\"187 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2018-02-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2018 21st Conference on Innovation in Clouds, Internet and Networks and Workshops (ICIN)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICIN.2018.8401580\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2018 21st Conference on Innovation in Clouds, Internet and Networks and Workshops (ICIN)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICIN.2018.8401580","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Destination-specific maximally redundant trees: Design, performance comparison, and applications
Resilient networks react to a failure by reconfiguring routing tables in a network-wide manner and utilizing fast reroute (FRR) in the meanwhile. FRR uses pre-computed backup paths to avoid the failure locally. The existing FRR approaches for IP networks and Software-Defined Networking (SDN) differ in regard to computational complexity, backup path length, additional forwarding state and required capacities. Maximally Redundant Trees (MRTs) are standardized by the IETF and are already available for IP and MPLS networks. In this paper, we propose destination-specific MRTs (dMRTs), an improved variant of MRTs, for protection of IP networks. This mechanism reduces the generated path lengths but increases the computational complexity of the MRT algorithm. We analyze and compare our proposal with simple MRTs and MPLS FRR. We show on a large and publicly available network data base that dMRTs provide significant shorter backup paths compared to simple MRTs and require less forwarding state and network capacities than MPLS FRR. Adapting an existing FRR mechanism for IP networks to SDN enables its deployment in hybrid-SDN networks. Thus, we outline implementation options of dMRTs in IP and MPLS networks and (d)MRTs in SDN. Finally, we discuss the computational overhead of our proposal in the context of (decentralized) IP networks and (centralized) SDNs.