Towards a dynamic adaption of capacity in mobile telephony networks using context information

As of today, components of radio access networks are operated in an always-on manner. Since network providers need to fulfill the traffic demands of all customers even during peak times to stay competitive, a significantly higher amount of network resources (i.e., bandwidth) is provided. By the end of 2010, Germany's mobile network providers operated more than 123.000 base stations which are available 24/7 — even when the provided resources are not required at this time. The total energy consumption of these base stations sums up to approx. 1455 GWh per year, inflicting not only high costs to the network operators, but also contributes to worldwide carbon dioxide emissions. Context data, which is already present in the different network entities can be utilized to model the state of radio networks allowing network providers to identify idle components. These components can then be reconfigured dynamically according to the actual demands on connectivity in the network. In this paper, we propose a Context Management Architecture, which is able to acquire and consolidate context from various components in radio access networks as well as additional external sources. The raw context data is sensed by special Context Collection Agents, which are installed on the devices the context data is originating from. It is then published to Context Managers, which provide functionality for storage, reasoning mechanisms, and provisioning of the context data. Context-aware applications (e.g., an application capable of optimizing the configuration of radio networks) can access the refined context via appropriate interfaces for further processing.