ParDBMS

Funding/Industry: Research Committee of the Aristotle University

Duration: 12 months (1/10/1992-30/9/1993)

Budget: 0.8 Million GRD

Category: National Projects

The main aim of this project was to develop a parallel Database Management System with a query language based on fuzzy logic. This was achieved by the interconnection of Logic Programming with Databases and the parallel processing of the latter. Modern applications ask for large volumes of data and complex query languages. The efficient management of such Databases is limited by the size of information that can be directly stored into main computer memory and by the search and retrieval time.

For the satisfaction of the above demands, the following solutions are generally adopted:

  • The extension of the relational model by using predicate calculus to represent data and to describe new logic query languages.
  • The evaluation of queries posed to the Database in parallel, using a multi-processing machine, in order to reduce the response time of the Database.

During this project, the above topics have been studied and the following solutions were proposed:

  • The development of a Database Management System on top of a Prolog system, since the latter is declarative and unifies data with knowledge and queries.
  • The use of a multi-processing machine with transputers in order to execute the above system in parallel.
  • The access of large Databases directly from the secondary storage devices (hard disks or floppy diskettes, for example), instead of using the main memory only.

The parallel deductive database system system is based on the top-down evaluation of logic programs. Parallelism is provided at the rule level, by transforming the query AND/OR tree into Disjunctive Normal Form. The clauses of the transformed formula are executed independently in parallel, on a transputer multi-processor machine, using the processor-farm algorithm. Both main-memory consultation and direct disk access have been implemented and tested. The measurement of the system performance showed speed improvement over the sequential Prolog interpreter, for large rule bases, but also exhibits implementation-dependent drawbacks that cause underlinear speed-up.

In the project report we also introduce fuzzy logic into the system to represent uncertain knowledge and data, in order to reduce the solution space and to improve the query language.

Team

  • Panayotis Linardis, Coordinator
  • Prof. I. Vlahavas
  • Associate Prof. Nick Bassiliades
  • Christophe Maciazek