Traffic-Driven Dynamic Spectrum Auctions

Wireless growth has been limited by the shortage of radio spectrum. While the spectrum assigned to legacy technologies remain unused, new prominent technologies such as Mesh/WiFi networks are forced to crowd into a small unlicensed band, suffering from significant interference and degraded performance. Using economic incentives, dynamic spectrum auctions redistribute spectrum to make it available to new technologies while providing financial benefits to legacy owners. In this paper, we investigate the performance of dynamic spectrum auctions under traffic dynamics. Using measured traffic traces from deployed WiFi access points, we evaluate the advantages and artifacts of dynamic auctions over plain channel sharing, and investigate the impact of bidding formats and auction intervals. Our results show that short-term dynamic auctions with traffic-aware bidding can significantly improve system throughput and provide bidders with cost-effective spectrum usage.