Object-Oriented Databases (OODBs) have gained wide attention in the past few decades because its greater modeling power on structurally complex data compared to relational database systems. These systems, benefited by the object-oriented approach, require also a high quality of performance on data manipula-tion. Indexing is one common technique to increase the performance of OODBs. Since most of these systems are static in nature, dynamic operations like class versioning and role-playing or intelligent operations like query relaxation and specification are not supported by traditional indexing schemes. In order to support these dynamic and intelligent query operations, a more sophisticated indexing scheme is needed. This thesis proposes the double signature indexing scheme (DSIS) for a...[ Read more ]

Object-Oriented Databases (OODBs) have gained wide attention in the past few decades because its greater modeling power on structurally complex data compared to relational database systems. These systems, benefited by the object-oriented approach, require also a high quality of performance on data manipula-tion. Indexing is one common technique to increase the performance of OODBs. Since most of these systems are static in nature, dynamic operations like class versioning and role-playing or intelligent operations like query relaxation and specification are not supported by traditional indexing schemes. In order to support these dynamic and intelligent query operations, a more sophisticated indexing scheme is needed. This thesis proposes the double signature indexing scheme (DSIS) for a dynamic object-oriented database programming language with role extension, called DOOR. The scheme supports retrievals of a large range of queries including class relation queries, nested queries as well as queries which only exist in dynamic environment like qualification and versioning queries.

In this thesis, the types of queries supported by the indexing scheme are addressed. Besides, the index structure of the indexing scheme is described. The dynamic, intelligent as well as static operations which are supported by the indexing structure are also stated. In addition, the processes on how the new indexing mechanism handle different kinds of queries are demonstrated. The maintenance of the index structure in the dynamic environment is also presented. Finally, analytical cost models for the indexing scheme is developed and the estimated retrieval and storage cost are formulated.