Channel dynamic adjustment in data broadcast

Abstract

Data broadcasting has become the preferred method to dispense data to a large number of mobile users. Current researches on on-demand data broadcast mainly propose algorithms based on a single broadcast channel or fixed multi-channel, i.e., fixed channel model. As a result of the dynamic diversity of data characteristics and client demands, the fixed channel model faces significant challenges in parallel broadcast diverse data. Further, the dynamic adjustment of the broadcast channel (dynamic channel model) based on client requests is favorable to service quality because it determines the number and sizes of channels that adapt to client demand in real-time. However, the dynamic channel model has not yet been thoroughly investigated for on-demand wireless data broadcasts. Accordingly, in this paper, a channel dynamic adjustment method (CDAM) is proposed. The innovations behind CDAM lie in three aspects. First, a data item priority evaluation and selection algorithm (S-RxW/SL) is proposed for evaluating the priority of data items and selecting the high priority data items to be considered in a broadcast cycle. Second, a weight and size average cluster algorithm (WSAC) is proposed for mining data item characteristics and clustering them. Third, based on the clustering results of WSAC, a channel splitting and data allocation algorithm (CSDA) is proposed for dynamically splitting the channel and allocating data items to the corresponding sub-channel. We compare the proposed method with some state-of-the-art scheduling methods through simulation. The theoretical findings and simulation results reveal that significantly better request loss rate (LR) can be obtained by using our method as compared to its alternatives.