Pull vs push: a quantitative comparison for data broadcast

Advances in wireless technology have resulted in an emerging broadcast-capable infrastructure. A major debate for such infrastructures is whether we should use push or pull to support large client populations. To date, push was suggested to provide a scalability that is not possible with pull. In this paper we conclude otherwise, based on simulation and experimental prototype results. A major contribution of this paper is to show that pull-based algorithms outperform push-based ones. Contrary to the conventional wisdom, this is true even for environments where request processing and scheduling overheads are a significant factor.