Efficient power -aware data access in mobile environments

In the last decade, there has been tremendous growth in both wireless communications and the Internet. These two technologies are making ubiquitous computing possible. We can envision the scenario of the near future where wireless data access will be readily available to the public anywhere and anytime. Before this vision becomes a reality, however, some issues still need to be resolved. Because of the physical limitations of the wireless network and devices, such as limited bandwidth, battery power, and computation power, it is much more important to design efficient data access schemes in wireless environments than in wired environments. This thesis addresses these issues by designing data access schemes in different wireless environments to improve their performance.;We first address the data access issue in single-hop based mobile environments, where mobile nodes access data through a base station that is one hop away. In this environment, the base station usually adopts the broadcast technique to disseminate data to mobile nodes. As data are available on the broadcast channel, mobile nodes may prefetch them to reduce query latency. However, prefetch consumes valuable battery power and hence should be performed carefully. Therefore, we propose a value-based prefetch scheme to determine which data to prefetch. Then adaptive prefetch schemes are designed so that the prefetch can be done while adapting to mobile nodes' power levels to achieve a balance between query latency and power consumption.;When there is no pre-installed base station, mobile nodes may form a multi-hop based wireless network and forward packets for each other. As long as one mobile node maintains a database or connects to the Internet, other mobile nodes can still access data through multi-hop links. In this environment, we propose to use two techniques to improve data access efficiency: cooperative caching and data replication.;Cooperative caching fully exploits the potential of caching by serving requests for other mobile nodes besides the cache owner. In mobile environments, a cooperative caching scheme should adapt to the mobile network where mobile nodes have limited computation and communication power, and where the network topology changes frequently. We propose several schemes that are suitable for mobile environments. These schemes are able to reduce query latency and power consumption without incurring high overhead.;Multi-hop based mobile environments have a high link/node failure rate, and thus the data access may be affected when the requested data item is not accessible because of network partitions. Data replication schemes are studied in this thesis to address this issue. Existing schemes usually focus on either reducing the query delay or improving data accessibility, but fail to consider the tradeoffs between the two performance metrics. The data replication schemes proposed in this thesis show that these two metrics are closely related and balance the tradeoffs between them.;Although this thesis studies different mobile environments, the focus is on efficient data access to reduce power and other resource consumption. We show that through intelligent use of the local cache, either by prefetching, cooperative caching, or data replication, mobile nodes are able to access data more efficiently. We hope that these research results can help build a better wireless environment for future mobile users. We also hope that other researchers find our analytical and simulation methods helpful so that new and more exciting results can be generated.