Data delivery techniques in wireless environments are important for applications such as news feeds, monitoring, military operations, and civilian outdoor applications. In such environments, clients are typically portable battery-powered devices with very limited computation and communication capability and low power reserves. Also, communication in wireless networks are very expensive. Efficient delivery of application-specific data and conserving device power are important design issues. In this thesis, we study several data delivery schemes suitable for different wireless communication environments.; First, we propose a data broadcasting model for asymmetric wireless communication environments where a central server maintains a collection of documents, and constructs adaptive broadcast schedules, based on randomized feedback from clients, to minimize the average waiting time across the clients. Clients in this model do not target individual data items, and are satisfied with responses that are sufficiently close to the desired information. We also show how clients may conserve energy by selective tune-in.; We next present a data delivery model for time-sensitive data in wireless sensor networks, in which all sensor nodes send sensor readings to a single base station, each associated with a deadline. We propose a load balanced routing scheme and a data packet grouping scheme that considerably reduce data transmissions and network congestions, increase the likelihood that data meet their deadlines, and save significant powers.; Finally, we consider location-based routing, which has become an efficient and scalable routing solution for wireless mobile ad-hoc networks. Such protocols require location information. We propose a new location service scheme that uses a hash-based virtual hierarchy to maintain the location information at mobile nodes and solves location tracking and querying issues efficiently. Our location service scheme is tunable to the network workload so that power savings can also be achieved.
展开▼
