Issues in indexing techniques for object-oriented databases

Indexes are crucial in database systems to expedite the evaluation of queries that retrieve a small subset of a large database. Indexing techniques have been widely investigated in the framework of relational, and network databases and several organizations have been proposed. The most common way to implement an index is by a B-tree structure or hashing. However, the novel features of object-oriented (and semantic) models pose some requirements toward conventional indexing techniques in order to efficiently support the types of queries that are possible in databases based on these advanced data model. In this paper, we discuss some issues concerning advanced indexing techniques, and we report results and point out open research directions.