Extending and configuring the PRIMA database management system kernel

Abstract

The authors give an overview of the mechanisms used in a specific DBMS (database management system) kernel, PRIMA, to easily extend and configure the system in advance, as well as on-line. They discuss four kinds of extensions: (1) schema extensions, that is, on-line addition and removal of object types and attributes without nasty data reorganization; (2) data model extensions, which allow for the introduction of new data types and operations for existing databases; (3) algorithmic extensions, which offer the possibility of improving the performance of a running system by the addition of new algorithms for existing operators and components; and (4) storage structure extensions, which can be used to organize already existing data (possibly redundantly) in newly integrated types of storage structures supporting the specific access characteristic of an application more efficiently.<>