Seamless MANO of multi-vendor SDN controllers across federated multi-domains

Computer Networks(2021)

引用 15|浏览14
暂无评分
摘要
The upcoming 5G networks promise to provide end-to-end network service delivery that spans across Cloud-Network federated multi-domains while providing monolithic service guarantees. By adopting a Cloud-Network federated multi-domain approach, heterogeneity raises a number of challenging Management and Orchestration (MANO) perspectives, especially concerning the need to deal with SDN controllers of multi-vendor approaches that lack a standard Northbound API. In this work, we tackle the issue of providing a seamless MANO of SDN Controllers running in Cloud-Network federated multi-domains. Owing to the weaknesses and limitations of the related works, we propose the WAN Infrastructure Manager Agnostic (WIMA) that enables a seamless and vendor-agnostic MANO abstraction to run on top of federated SDN multi-domains. The WIMA provides a common Northbound API for external triggering, maintains a global topology view of the federation as a whole, and deals with each SDN Controller directly by deploying an ontology-based scheme for efficient Northbound API mapping. The effectiveness and performance impact of WIMA was assessed in an emulated testbed with homogeneous (“Hom”) and heterogeneous (“Het”) multi-domain SDN control-planes, along with a varying density of active Tenants which simultaneously makes flow stress data connections during an experimental time of 900 s. The obtained results reveal that WIMA’s MANO abstraction system is able to connect around 52.66% (“Het”) and 86.87% (“Hom”) more end-to-end data flows across the federated SDN multi-domains while adding greater agility 52.72% (“Het”) and 85.27% (“Hom”) than the rival Baseline solution. Thus, the WIMA’s central logic has proven to be a suitable and feasible means of ensuring the MANO framework’s efficiency atop the multi-domain SDN Controllers within a Cloud-Network federation while optimizing the operation time.
更多
查看译文
关键词
Software-defined networking,Orchestration,MANO,Agnostic
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要