LearnQoS: A Learning Approach for Optimizing QoS Over Multimedia-Based SDNs

• Main Authors: Al-Jawad, Ahmed, Shah, Purav, Gemikonakli, Orhan, Trestian, Ramona
• Format: Conference Proceeding
• Language: English
• Subjects: Decision making; Learning (artificial intelligence); Loss measurement; Monitoring; Quality of service; Reinforcement Learning; Routing; Software Defined Networks; Streaming media
• ISSN: 2155-5052
• DOI: 10.1109/BMSB.2018.8436781

As video-based services become an integral part of the end-users' lives, there is an imminent need for increase in the backhaul capacity and resource management efficiency to enable a highly enhanced multimedia experience to the end-users. The next-generation networking paradigm offers wide advantages over the traditional networks through simplifying the management layer, especially with the adoption of Software Defined Networks (SDN). However, enabling Quality of Service (QoS) provisioning still remains a challenge that needs to be optimized especially for multimedia-based applications. In this paper, we propose LearnQoS, an intelligent QoS management framework for multimedia-based SDNs. LearnQoS employs a policy-based network management (PBNM) to ensure the compliance of QoS requirements and optimizes the operation of PBNM through Reinforcement Learning (RL). The proposed LearnQoS framework is implemented and evaluated under an experimental setup environment and compared with the default SDN operation in terms of PSNR, MOS, throughput and packet loss.