Mobility-aware Management of Internet Connectivity in Always Best Served Wireless Scenarios

The widespread availability of portable devices with multiple wireless interfaces, e.g., IEEE 802.11, WiMAX, Bluetooth, and/or UMTS, is leveraging the potential of novel supports to seamlessly and automatically select the proper connectivity technology to exploit at any time for any node and any running application. That selection should be context-dependent and take into account several aspects, at very different abstraction layers, from application-specific bandwidth requirements to expected client mobility, from user preferences to energy consumption. We claim the suitability of mobility-aware middle-wares to relieve the application logic from the burden of determining the most suitable interface and connectivity provider for each client/application at service provisioning time. In particular, the paper motivates the need for novel context indicators, e.g., client/connector relative mobility, and describes effective lightweight solutions to estimate them flexibly, depending on dynamically introduced evaluation metrics. The paper presents primary architecture and implementation guidelines to build such a novel middleware solution. The proposed middleware has been experimentally validated and the reported performance results demonstrate the feasibility of the approach: it achieves accurate estimations of node mobility and consequently per-forms connection establishment/selection with very limited overhead.