A data repository for fine-grained adaptation in heterogeneous environments Calicrates Policroniades, Rajiv Chakravorty, and Pablo Vidales   View Paper (PDF)   Return to Location awareness and moving objects Abstract In this paper, we present DATOM -- an extended structured data storage system for mobile data management. We propose cooperation of networks and applications in order to adapt to the limitations imposed by heterogeneous environments. We show how adaptation can be achieved to address the high variability in link-layer characteristics as typically seen in hybrid wireless networks.We propose a novel data model for DATOM, that departs from the traditional view of files as monolithic objects, and breaks them into elements to enable fine-grained adaptation. We incorporate metadata such as access patterns and priority to individual elements in our data model. Content adaptation of individual elements is controlled by application-specific policies that are applied according to present and future network conditions. Using an integrated LAN-WLAN-GPRS testbed, we propose to evaluate a set of adaptation strategies for mobile data management that can help in improving the overall user experience. ©2005 Association for Computing Machinery