To counter the inefficiencies of the current spectrum usage, regulatory bodies, all over the world, are exploring ways to deregulate the spectrum market by allowing flexible dynamic spectrum access (DSA) in a broad range of spatio-temporal scale. Recent advances in radio technology have given an impetus to this trend. For DSA to fulfill its promise of economic and societal impact, wireless services based on DSA must be commercially successful, and a tangible spectrum market must evolve that can be supported by technology. This research project will build a realistic DSA architecture for cellular networks supported by appropriate market mechanisms in an integrated fashion that is both technically and economically viable and efficient. This is a truly trans-disciplinary approach spanning the fields of wireless networking and systems, algorithmics, economics, simulation and modeling, which leads to a deeper understanding of the dynamics of the spectrum market by (i) realistic modeling of various market entities (i.e., buyers, sellers, and the market mechanisms), (ii) dynamic spectrum demands and bids based on innovative and realistic population dynamics models, and (iii) new and robust market clearing mechanisms with provable performance guarantees. The results will be validated using large-scale simulations, and experiments on a prototype test bed with reconfigurable radio hardware. In addition to fostering new topics in trans-disciplinary education, this project will offer insights into market driven spectrum sharing, provide useful tools for policymakers, and ultimately guide spectrum policy decisions in DSA technology. This will, in turn, open up new business opportunities in the use of wireless spectrum.
The research funded under this grant aimed to explore a new radio spectrum access architecture in the context of cellular wireless networks, namely that of Coordinated Dynamic Spectrum Access (CDSA) which relies on the notion of a network resident spectrum broker managing access to spectrum resource. The Bell Labs PI, a pioneer in the concept of CDSA, worked with his academic collaborators and supported students as Bell Labs summer interns to address a few salient aspects of CDSA. Our research focussed on three broad areas: (1) Develop techniques and tools to model dynamic spectrum demand and bids using analysis of real cellular traffic data. (2) Design fast algorithms for real-time spectrum allocation as well as network demand management to optimize spectrum utilization and other network performance metrics such as energy used. (3) Develop a prototype testbed using reconfigurable radio hardware, to demonstrate end-to-end functioning of our techniques, and to gain realistic experience in operation of CDSA based wireless networks. Our research project has been a great success on many fronts: (1) During its course, it provided high quality internships and state-of-the-art training to seven graduate students . The work supported under this grant also became part of the Ph.D. thesis of two of these seven students. (2) We published 7 research papers, out of which one paper reporting our prototype system that integrates statically allocated traditional spectrum with broker controlled spectrum in a small cell won Best Technical Paper Award at IEEE DySPAN 2012. Our paper on cellular traffic analysis published in IEEE Infocom 2011 conference has been widely cited (145+ times as per Google Scholar) in a short span of 3 years and as such it has provided much needed insights to other research efforts on resource management in cellular networks. (3) Our work also resulted in filing of two US Patents, one of which is already granted. Our research made significant progress in developing technologies required for exploiting spectrum in a shared fashion and gaining greater understanding of spectrum use in state-of-the wireless networks. A few of the signficant results of our work are as follows: (1) Our work on understanding spatio-temporal characteristics of cellular networks provides ways to model spatial and temporal use of spectrum channels and therefore, combining with a human behavior model, our insights can be used to predict spectrum demands essential for market driven trading and spectrum management. (2) Our work also showed that we can use spectrum allocation and management as a dial to control energy consumption in cellular networks and create an energy market where spectrum can be used as a resource for trading energy costs. We showed that using lower spectrum bands during time periods of low network demand, a high percentage of network basestations can be turned off to signiificantly decrease energy consumption in the networks. (3) We also showed that defering delivery and consumption of high bandwidth content (e.g.: such as video) to duration when network is less loaded, the network capacity utilization can be made significantly uniform, reducing peak capacity and spectrum requirements. (4) Our end-to-end testbed conclusively showed how a dynamic spectrum broker can allocate channels to the network of base stations and how such channels can be used in cellular small cells to minimize interference from macro-cells in which they are embedded. Our work on such dual technology small cells -- that support traditional fixed licensed or unlicensed channels with dynamic channels won the best paper award. (5) We also showed that certain spectrum management tasks in cellular networks can be implemented using a more scalable, distributed approach also known as "Self-management" to minimize "single point of failure" in the form of spectrum broker. In recent years, we have witnessed growing activitiy and interest world wide in dynamic access to spectrum using spectrum databases. In the US, Federal Communication Commission's Digital TV Whitespace ruling and ongoing initiative to release 3.5 GHz spectrum for shared use advocate Spectrum Broker/ Database based approach. Our research efforts were ahead of the times in predicting this approach and as such our research results help provide evidence of technical feasibility of such an approach. The technologies researched and prototyed in this research grant represent early steps in building future ultra-broadband wireless networks essential for commercial and first responder public safety communication needs. In fact, these technologies have potential to revolutionize business of deploying and operating wireless networks and services.
