Dynamic Multipath Routing Mechanism for Multimedia Data Flow Scheduling over Software Defined Networks

Wu Jiawei,Qiao Xiuquan,Lu HuiJuan
DOI: https://doi.org/10.1109/dsn-w54100.2022.00040
2022-01-01
Abstract:Due to the explosive growth of big data, driven by the pervasiveness of real-time multimedia applications, Software-defined Networks (SDNs) are expected to fully handle multimedia services with various quality of service (QoS) guarantees. Unlike traditional data applications (e.g., e-mail and file transfer), multimedia applications require diverse QoS guarantees to satisfy the user’s expectations. Currently, service providers still have difficulties in providing high quality video streaming due to the difficulty of scheduling efficiency and of end-to-end data delivery in the existing multipath algorithms. Moreover, existing multipath algorithms suffer from a significant transmission cost. To solve these problems, a Dynamic and Adaptive Multi-path Routing algorithm with respect to multiple constraints (DAMR) over SDN is proposed in this paper. The proposed approach leverages the Lagrangian Relaxation Algorithm (LRA) to find a feasible path that satisfies the desired QoS requirements. The DAMR takes several QoS routing constraints, such as end-to-end delay and bandwidth, into account in path selection. Compared with existing routing proposals, the DAMR makes efficient use of the latest global network state information achieved by the OpenFlow controller and it dynamically calculates the optimal routes according to the real-time status information of the link. Theoretical and simulation results show that our proposed DAMR algorithm is better than SPT and MCMPT in terms of packet loss ratio, throughput and end to end delay.
What problem does this paper attempt to address?