Heterogeneous Query Language Techniques for Data Integration
The project aims at developing next generation query language
processing technology for integration of heterogeneous external data
sources. The data sources can, e.g., be relational databases,
object-oriented databases, data exchange formats such as STEP/EXPRESS
or XML, conventional files, or data-producing programs. Integration
of different kinds of data sources will require a high-level {\em
object-oriented multi-database language}. Central problems here are
definition, processing, and optimization of object-oriented
multi-database query language constructs. Object-oriented queries and
views will be used for handling {\em semantic heterogeneity} between
the data sources.
In our research, we are developing a light-weight database mediator
system. The system is able to integrate existing legacy database
systems and provide the user with a location transparent and
semantically consistent view of data stored in these databases. The
following four mechanisms have been identified as important for such
mediation facilities: object-oriented (OO) database view subsystems,
interdatabase communication protocols, query decomposition algorithms,
and OO mechanisms for resolving semantic heterogeneity conflicts
between views. We have developed a heterogeneous database view
mechanism in the AMOS light-weight database system. The results are
published in a journal, a conference, and a workshop. We are
currently working on efficient execution of queries to the
heterogeneous views and query decomposition. The project resulted in a
PhD Thesis for Vanja Josifovski.
Graduate student: Vanja Josifovski (Ph.D. 1999) (moved to Uppsala universitet).
Supervisor: Tore Risch.
