A near optimal scheduler for on-demand data broadcasts

  • Authors:
  • Hing-Fung Ting

  • Affiliations:
  • Department of Computer Science, The University of Hong Kong, Hong Kong

  • Venue:
  • CIAC'06 Proceedings of the 6th Italian conference on Algorithms and Complexity
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

In an on-demand data broadcast system, clients make requests for data such as weather forecasts, stock prices and traffic information. The server of the system broadcasts the requested data at some time, and all pending requests on this data are satisfied with this single broadcast. All requests have deadlines. The system can abort the current broadcast for more valuable requests and a preempted broadcast may be restarted from the beginning later. In this paper, we design and analyse online scheduler for scheduling broadcasts in such system. The best previously known upper and lower bounds on the competitive ratio of such schedulers are respectively $\Delta + 2 \sqrt{\Delta} + 2$ and $\sqrt{\Delta}$, where Δ is the ratio between the length of the longest and shortest data pages. In this paper, we design a scheduler that has competitive ratio $\frac{6\Delta}{\log \Delta}+O(\Delta^{5/6})$. We also improve the lower bound of the problem to $\frac{\Delta}{2\ln \Delta}-1$, and hence prove that our scheduler is optimal within a constant factor.