Date of Award


Publication Type

Master Thesis

Degree Name



Computer Science

First Advisor

Ezeife, C.


Computer Science.




In intranets, extranets and internet applications, data are by nature complex and distributed over different sites. Object-oriented database management systems meet the requirements of these applications. It offers complex structures, object identity, inheritance between classes and extensibility to capture complex data. A distributed database system partitions large and complex data into smaller pieces and allocates them at different sites to enhance application performance by reducing data communication and replication costs. The design issues of distributed database system require solving several interrelated problems: data fragmentation, allocation and optimization. There are three types of fragmentation---horizontal, vertical and hybrid. Horizontal fragmentation of a class keeps all attributes and methods of the class but some instance objects in each horizontal fragment. In other words, a horizontal fragment is a subset of class extent or instance objects. Application queries, query access frequencies, instance objects, and object database schema including class composition hierarchies and class inheritance are used as input to generate these fragments. When there are major changes in these input over time, the performance of the distributed object-based system degrades and requires re-fragmentation. The re-fragmentation is started from scratch with static fragmentation approach using all input data (old and changed part). In this thesis, we propose a new algorithm called Incremental Object Horizontal Fragmentation (IOHF) for distributed object-oriented database systems. This algorithm uses the changed part of input data and previous fragments to define new fragments more quickly, saving system resources and making data at distributed sites more available for network and web access. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis2002 .D49. Source: Masters Abstracts International, Volume: 42-01, page: 0256. Adviser: Christie Ezeife. Thesis (M.Sc.)--University of Windsor (Canada), 2002.