Issue No. 10 - October (2007 vol. 6)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TMC.2007.1044
The emerging proxy-based wireless Content Delivery Networks (CDNs) should to be designed to download hugesize files over fading-affected channels. However, from a radio resource management point of view, several basic problems need to be still solved for such wireless delivery systems to operate efficiently. Specifically, due to the fading nature of the downlink channel, a still open basic problem is to design optimal energyallocation (e.g., scheduling) policies that minimize the requested download-time when constraints on the total available energy and peak-energy are simultaneously active. In this contribution, this problem is solved for application scenarios where the downlink channel is slotted and continuous-state, the carried out traffic is elastic and the resulting conveyed throughput is measured by any desired increasing concave rate-function. Specifically, the optimal energy-allocation policy minimizing the download-time is computed in closed-form and its performance is compared against that of a basic On-Off heuristic energy-scheduler on some Rayleigh-faded Multi-Antenna delivery systems of practical interest. The carried-out performance comparisons point out that the presented optimal policy typically may outperform the heuristic one up to two-orders of magnitude, specially when the delivery system is strongly energy-limited.
Minimum download time, elastic traffic, energy-scheduling policies, CDNs, Multi-Antenna downlink channels, convex Calculus of Variations
Mauro Biagi, Cristian Pelizzoni, Nicola Cordeschi, Enzo Baccarelli, "Minimization of Download Times for Large Files over Wireless Channels", IEEE Transactions on Mobile Computing, vol. 6, no. , pp. 1105-1115, October 2007, doi:10.1109/TMC.2007.1044