The first serious work of the MODIS team was concerned with design and development of GNU SQL Server (late 80s-middle 90s). The project was supported by Free Software Foundation (www.fsf.org). The main goal of the project was to create free multi-user fully functional relational DBMS conformed to the standards of the SQL query language. To perform the project a big research work was done. The methods of cost-based query optimization were analyzed and improved. The algorithms and protocols for transaction management, logging, and recovery were investigated. The best known techniques (especially from System R publications (http://www.mcjones.org/System_R) as well as some their improvements were used. The most interesting results of the research are as follows:
- cost-based query optimization with non-uniformed column values distribution
- predicate-based locking schema for transaction coordination
- separate logical and physical logging and recovery
The system had been implemented and may be downloaded here or on sites of Free Software Foundation.
At the end of 90s the main R&D area of the MODIS team was CORBA-based technology for distributed information systems. To provide a test-bed for future research, and to get a practical experience with CORBA the own OMG-compliant ORB had been implemented. It was called C++ ISPORB. The system can be downloaded for free here.
More practical (and commercial) aspects of the work were related with cooperation with Nortel Networks company. By contract with the company Protel-2 ORB was developed for Nortel Networks Open Node project. Protel is the proprietary programming language of the Nortel. Object request broker for Protel-2 language is an example of broker, implemented for usage in the special environment of telecommunication module.
From the beginning of 2000s the MODIS team is concentrated on XML-based technology for heterogeneous data integration. In cooperation with international company Advanced Technical Services ATS (HITECH-FZE) the project BizQuery is under development. The BizQuery suite is founded on two major technologies - XML and UML - and provides facilities for enterprise data modeling and integration and for easy access to integrated data that can be XML or relational data. The main technical features of BizQuery are as follows:
- Support of open standards ensures ease of programming and effortless integration with the existing applications (XML, XML name spaces, XSLT, XQuery, OMG XMI, OMG UML, etc.)
- Accessibility via a two level interface offers two ways to query data: in terms of XML as well as in terms of the UML-based model
- XQuery support appeals to the customer by XQuery-s expressiveness
- Automatic but customized generation of three kinds of the end user query interfaces (i.e. forms, graphic map for querying with the support of a user friendly UML notation, catalogs)
- Easy interaction from any language and environments using SOAP. This open web service interface enables easy integration for your applications written in any language (Java, VB, Perl, C/C++, C#, Python)
More information about the BizQuery project is available here.