A SDN-based On-Demand Path Provisioning Approach across Multi-domain Optical Networks

Md Israfil Biswas, Philip Morrow, Mamun Abu-Tair, Sally McClean, Bryan Scotney, Gerard Parr
2018 2018 IEEE 20th International Conference on High Performance Computing and Communications; IEEE 16th International Conference on Smart City; IEEE 4th International Conference on Data Science and Systems (HPCC/SmartCity/DSS)  
The interconnection of remote datacentres with optical networks are emerging use cases and such orchestration of multi-domains require the design of new network control, management, and orchestration architectures. Such heterogeneity needs to adopt end-to-end services like on-demand path provisioning. It is acknowledged that such scenarios are more complexed and have fundamental limitations in terms of high performance and delay. To address these issues, and as a means to cope with the
more » ... y growth, research in this area is considering the concept of Software-Defined Network (SDN) orchestration for multi-domain optical networks to coordinated the control of heterogeneous systems. This paper presents a SDN path provisioning approach across Multi-Domain Optical Networks. The aim is to develop an efficient on-demand path provisioning platform in a software defined optical network at the control plane to dynamically manage the network's load, especially in emergency scenarios. The proposed distributed system architecture will help to solve the longstanding problem of inter-domain path provisioning. Our proposed architecture is implemented and validated in a control plane testbed to validate the approach. The paper also evaluated the factors such Quality of Service (QoS) of the network deployment associated with delay or control overhead. Our results show that the method will reduce additional delays in a multi-domain optical network, where high capacity and low latency are requirements for data-intensive applications and cloud services. The proposed method also maintains the total number of flows as low as possible to make the algorithm fast and reduce overheads.
doi:10.1109/hpcc/smartcity/dss.2018.00150 dblp:conf/hpcc/BiswasMAMSP18 fatcat:5yq3qo4lenenreynu7y46r4cei