An indexing model for object-oriented database systems

In an OODBS (object-oriented database system), access to the persistent objects (i.e. objects stored in secondary memory) dramatically influences the system efficiency because of their size and data structure. In a relational DBS, special data structures, called indexes, are used for speeding up the process of locating data according to requirements of specific applications. It is argued that indexes in OODBS should be used for accelerating the process of locating data and for improving the reconstruction of objects in main memory. Since the structure of objects can be complex, other objects, which are part of the retrieved object, can be involved in the read operation. An examination is made of the specific tasks that an indexing mechanism should perform within an OODBS. A uniform view is given of different indexing strategies, from a logical point of view. A simple data model for objects is described. The indexing abstract model and three indexing techniques for OODBS are also presented.<
>