Broadcasting Data for Real-time Multi-item Requests in Multi-channel Environments

Mobile computing pervades our daily lives as millions of people are now using some kind of mobile device with a wireless interface to access information for business or personal use. Demand is expected to increase for time-critical information dissemination services such as mobile stock trading and location-based services. This project proposes a two-phase processing methodology for developing a robust and responsive on-demand data broadcast system that produces high-quality broadcast schedules for these applications in multi-channel environments. The research issues include system timeliness and predictability, effective sharing of data, flexible channel allocation, and system overload management. First, the researchers will design a novel scheduling algorithm that considers request timing requirements and data sharing among multi-item requests to schedule incoming requests to an admission control module that determines whether there is sufficient broadcast bandwidth among different channels to admit the requests. Second, they will design a robust channel allocation mechanism to allocate data items of the admitted request into channels to meet its deadline. The real-time performance and effectiveness of bandwidth usage of the proposed algorithms will be evaluated by performing a series of extensive simulation experiments.