Implementation of spectrum micro-trading for mobile operators in the spatial dimension

Radio spectrum is a limited natural resource. Most spectrum resources are assigned in the form of licenses, but the actual usage of most spectrum bands is limited. Market mechanisms to trade unused spectrum resources is a promising approach to mitigate the spectrum crunch. An extremely flexible approach to this, the spectrum micro-trading pixelation model, has been proposed to trade spectrum on the micro-scale in frequency, time and geography. In this paper we implement this model in a simulator and evaluate its performance in the spatial dimension using different pixel sizes and auction models. Some surprising findings are presented; the auction that assign spectrum to the bidders willing to pay the most does not provide service to most customers, does far from optimally maximize spectrum utilization, and gives far lower profit to the spectrum broker facilitating the market. Furthermore, we demonstrate the critical tradeoff between spectrum exploitation efficiency, allocation overhead and geographical pixel sizes in the market.