Towards Efficient Multi-Channel Data Broadcast for Multimedia Streams.
Xiaofeng Gao,Ailun Song,Luoyao Hao,Junni Zou,Guihai Chen,Shaojie Tang
DOI: https://doi.org/10.1109/tpds.2019.2908904
IF: 5.3
2019-01-01
IEEE Transactions on Parallel and Distributed Systems
Abstract:Multi-channel data broadcast attracts increasingly focus in recent years. For the complex and large multimedia data like images, audios, and videos etc, multi-channel data broadcast is a promising approach to mitigate various limitations of data dissemination in mobile environment, such as narrow bandwidth, unreliable connections, and battery limitation. However, existing data broadcast schemes are inefficient for employing multiple channels. In this paper, we present five novel multimedia data broadcast schemes (SDAA, ISDAA, LDAA, AEA, and COA) specifically designed for wireless multichannel communications. The key idea behind those schemes is to utilize SVC (Scalable Video Coding) to generate data segments with different qualities and implement indexing and channel assignment to minimize the expected waiting time for clients. In terms of waiting time minimization, we prove theoretically that SDAA is a 4/3-approximation algorithm and ISDAA is a 6/5 -approximation algorithm for global allocation, and AEA can achieve locally optimal solution. We show that integrating SDAA and AEA form a best schedule for practical applications. We provide numerical experiments by using a SVC dataset and a YouTube dataset to evaluate the system performance and prove the efficiency of our schemes. We also propose a multiple video broadcast framework and analyze its performance.