Database file systems
LE3. A278 2009
Bachelor of Computer Science
Since the time of the earliest computers, there has been a need to store and retrieve data. Whether this be on punch cards, tapes, magnetic or optical disks, or more recently solid state drives, the data must be stored. While hardware has evolved signicantly, the paradigm has not. Data is stored in files, which are stored in directories, which are then arranged into a directory tree. Organizing the directory tree is the end user's responsibility, and the file system is ignorant of any methodology to it. To overcome this limitation the file system must be made smarter, educated in the ways of data organization. Databases specialize in data storage, maintenance and retrieval, and are an excellent fit for the task. A file system which incorporates data organization techniques from databases would no longer be ignorant of the data stored within. Unfortunately, combining these two technologies is no easy task; there are many prob- lems which need to be overcome. One such problem is translating the information needed for a file system into something usable within a database. Rather than using the traditional directory tree, files will be tagged with metadata (some automatically generated, some user generated). Files will then be located via this metadata, which makes for a much more malleable approach to file organization. My research has focused on bridging the gap and assembling a blueprint according to which a database le system could be built.
The author grants permission to the University Librarian at Acadia University to reproduce, loan or distribute copies of my thesis in microform, paper or electronic formats on a non-profit basis. The author retains the copyright of the thesis.