An on-demanded data broadcasting scheduling considering the data item size

Abstract

Compared with conventional data broadcasting, on-demand data broadcasting is adaptive and real-time, which can better reflect the actual needs of mobile users. Current researches do not consider the attribute of data item size, and the constantly changing characteristics of data item size in on-demand data broadcasting is non-ignorable. This paper introduces the split strategies and backpacks theories into on-demand data broadcasting scheduling to deal with the inconsistencies of data item size, and proposes two scheduling models under different split strategies: (1) equal split scheduling model ES-LxRxW, which proposes the equal splitting strategy (ES) and a deadline adjust strategy. (2) Unequal split scheduling model US-LxRxW, which proposes the unequal split strategy (US) and two effective scheduling algorithms priority first (PF) and propriety and bandwidth first (PxBF). Extensive experiments shows that ES-LxRxW and US-LxRxW can both improve bandwidth utilization and dynamically adjust to the real-time situation of data item size, which takes into account data item size, bandwidth, cycle and scheduling priority of data item. The two proposed scheduling models could reach or outperform the other state-of-the-art scheduling algorithms without considering data item size in the performance of request drop ratio. US-LxRxW can also better reflect the real-time changes of data items than ES-LxRxW, and the proposed PF and PxBF algorithms can effectively improve the bandwidth utilization and reduce the split times.