Incorporation of Data Files into Semantic Databases

摘要

This project supported the investigation of ways to incorporate heterogeneous data files as well as information about these files into a high performance semantic database. The incorporation of these files will allow scientists to quickly search a diverse set of information in ways that are impractical using current databases. Our methods do not logically change the data files, allowing the programs that are currently used to generate, process and access these files to remain in operation. Towards this goal, we have enabled our database engine to store raw values of attributes of arbitrary length, e.g., a datum 2GB long, transparently fragmented and load balanced among many disks comprising the database with highly efficient access to any offset within this attribute value datum. We are developing a loading program that facilitates the import of data into our Semantic Object oriented database system, technology that streamlines access to arbitrary WWW data, and technology that allows us to easily design semantic access methods for other databases.