Approximation issues for soft QoS support in large-scale broadcast-based networks

Scalability is a major requirement in emerging systems. For instance, sensor networks demand extremely scalable solutions that can accommodate very high numbers of nodes. Broadcast-based networks are shown to provide a highly scalable infrastructure if the downlink can be used as a high-bandwidth shared-communication medium. For applications that have different timing requirements, though, it might be necessary to keep track of individual request information. In this paper, we study data broadcast server design for requests associated with specific response deadlines. In particular, we study FCFS, MRF and EDF for deadline-aware scheduling. Our results suggest that under an efficient implementation, deadline-aware MRF can outperform EDF.