Modern information dissemination services can dynamically produce real-time data that is valid and useful depending on users' on-demand data requests. Information servers must adopt a scheduling approach to maintain timely data access and system performance. Most previous studies on scheduling such data discuss on-demand data broadcasting and real-time task processing with various performance metrics. This paper exploits the attributes of request urgency, service productivity, and access fairness, and proposes an adaptive scheduling scheme for delivering on-demand, time-critical data objects over a data broadcast channel. In terms of request satisfaction and mean access time, performance results show that the proposed scheduling approach is amenable under dynamic on-demand time-critical requests, simultaneously achieving a low mean access time and request deadline miss rate.
