Optimizing video QoE for mobile eMBMS users in cellular networks
Zahran, Ahmed H.
Sreenan, Cormac J.
Institute of Electrical and Electronics Engineers (IEEE)
IEEE Evolved Multimedia Broadcast Multicast Service (eMBMS) is used in cellular networks to improve the utilization of scarce wireless resources in high user density service areas. However, eMBMS configuration involves interwoven decisions including which base stations (eNB) to synchronize to form Single Frequency Networks (SFN), which video qualities to be serviced, and how to distribute resources among different videos. These decisions should accommodate disparate channel conditions for eMBMS users and the impact of eNB's unicast-load in the service area. In this paper, we formulate eMBMS configuration as an optimization problem that maximizes the video QoE for users. Additionally, we present NIMBLE as an eMBMS configuration heuristic, guided by our optimization framework, to solve the problem in real-time. Furthermore, NIMBLE's design integrates elements to accommodate the dynamic nature of cellular networks resulting from changes in both user and network state over time. We developed a simulation testbed and performed extensive experiments to show that, in comparison to state-of-the-art schemes, NIMBLE can increase the average user throughput by 150% and reduce the bitrate switches by 75%.
Streaming media , Quality of experience , Bit rate , Resource management , Unicast , Optimization , Cellular networks
Khalid, A., Zahran, A. H. and Sreenan, C. J. (2020) 'Optimizing video QoE for mobile eMBMS users in cellular networks', IEEE Transactions on Multimedia. doi: 10.1109/TMM.2020.3021229
© 2020, IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.