Cost Effective Interference Management in Ultra-dense Hotspot Mobile Broadband Systems

Abstract: Rapid mobile data traffic growth is becoming in a reality and several forecasts expect that it will be continued in upcoming years. It is expected that significant indoor investment will be made not only by traditional operators but also by facility owners for their own purposes. A key challenge to such local network providers is provisioning ever-increasing mobile traffic demand at the current level of production cost per bit. A popular deployment strategy so far is deploying WLAN networks. While denser indoor deployment is foreseen, the interference from inside of a network as well as other neighboring operators can be a limiting factor for higher capacity. Tighter interference management will certainly provide higher efficiency in network and spectrum usage. Nevertheless, costs to allow fast information sharing among access points are necessary for advanced interference coordination. Moreover, managing interference across networks owned by different operators raises not only infrastructure cost but also the network interrelatedness which operators are typically reluctant for business independency. When taking into account the cost and barriers for interference coordination, it is still not so obvious that coordination in wireless broadband systems will be advantageous to operators.In this thesis, we address the operator benefit of downlink interference coordination in two aspects: 1) multi-cell coordination with no interference from neighboring operators, and 2) inter-operator coordination in shared spectrum. In order to deal with interference and cost tradeoff analysis, we explicitly develop a techno-economic analysis framework and reform a traditional cost model. Numerical results indicate that the economic benefit of the multi-cell coordination significantly depends on propagation conditions and average user demand level. A self-deployed WLAN network can be the cheapest deployment option in closed areas up to certain average demand level. Over the demand level or in open areas, advanced joint processing schemes in a cellular domain may be a viable solution. The drawback is that it requires extremely accurate channel state information at transmitters for practical usage. When inter-operator interferences is present, asymmetric cellular networks will be likely to appear due to business independency and selfishly compete to access spectrum with no or little network-level coordination. A network designed for more fairness with higher transmission power will have more benefit against the other counterpart. Although asymmetric competition lets operators unfairly utilize spectrum, sharing spectrum with reasonable geographical separation can outperform over static coordination, i.e., traditional spectrum split. Tight cooperation to maximize a common objective can further offer the performance benefit to both involved partners. However, the cooperation gain quickly diminishes as network separation and size increases because self-interference becomes more dominant.