Master student projects

The MonetDB experimentation platform creates a challenging environment to conduct research and development. The student projects range from development tasks for broader dissemination to projects geared at preparing for a PhD track.

Support for Scientific Data(ca 3 months)
Predominant scientific formats are HDF5 and NETCDF. They are produced by scientific experiments, simulations and consumed by GUIs. The target of this project is to extend the MonetDB kernel with a flexible scheme to make the content of these files accessible at the MAL and SQL level. It would greatly expand the applicability of MonetDB's user base.

MonetDB/tellme (6 months)
Databased query languages presuppose the user to be aware of the database content, both its schema and value distribution. A grand challenge is to squeeze SQL result sets into a summary to aid the user in formulation of better queries. (Java project)

MonetDB/Viewport (3-6 months)
The amount of numeric data stored in scientific- and commercial- database applications make searching it in an ad-hoc way with SQL rather cumbersome. For, the data distribution and correlations are often not known. A possible solution is to map the data into a visual world. Querying then becomes navigation in a virtual world. The target of this project is to explore a few visual datamining approaches on top of MonetDB. The target database is our skyserver implementation. (Java project)

MonetDB/SQLARRAY (ca 6 months)
The MonetDB/SQL database technology has high potentials in the area of scientific data management. However, the existing query language interfaces lack the abstractions needed, e.g. array support in SQL is still immature. The goal of this project is to identify enhancements to SQL to simplify the step for a scientist to use a DBMS. It is compared with the opportunities offered by a more functional-style language. (C project)

MonetDB/SPARQL (ca 6 months).
In this project we seek to provide a complete implementation of SPARQL, the embeding of RDF into SQL, on top of the MonetDB framework. We expect the kernel to be well-prepared to handle the large graph-based database querying in this application domain. (C project)

MonetDB Embedded (ca 6 months)
Embedded databases are widely used. Many small devices rely on hundreds of calls to organise user-specific behaviour. The challenge of this project is to realise a front-end for MonetDB embedded to study its performance consequences, e.g. compare it with Berkeley DB. The next step would be to improve the footprint of the code base, aiming for a version that could run on such small devices as a Gumstix part of applicances and robots. (C project)

MonetDB/LDAP (ca 6 months)
With emerging roaming of users, the need for a central maintenance and storage of user preferences increases. This kind of information is typically stored in a directory server, using the LDAP protocol. This project aims at creating such directory server based on MonetDB, to allow for very flexible addition or modification of properties in large environments, while still maintaining a high performance. The project can target as implementation within OpenLDAP as backend (via an MonetDB Embedded instance) or as stand alone application where MonetDB handles the LDAP client calls.

MonetDB SQL:2003 (ca 3 months)
The MonetDB SQL front-end is based on the 1999 standard and the SQL:2003 standard is approved. This development project is aimed at covering more functionality ground, e.g. the datamining functionality.

MonetDB in GIS applications (ca 6 months)
Geographical Information Systems challenge a database system in the support of both geometric data structures, algorithms, and management of large binary objects, e.g. satellite images. This project should scout the available approaches and extend MonetDB with functionality to become a viable alternative in this field.

MonetDB in Bio Informatics (ca 6 months)
Data-mining and searching in bio-medical information systems is an emerging field currently hardly supported by standard relational systems. In this project, the functionality of MonetDB should be extended to become a viable engine for scientific database, illustrated by an application in DNA sequence retrieval.

MonetDB Quality Assurance (ca 3 months)
The MonetDB project is a large sized adventure. Many hardware platforms, several sub-projects and a fair amount of code pose a severe challenge to its developers. Every night the system is tested on available platforms, and reports are sent out to the developers. However, this is not enough to avoid (compatibility) problems or solve them. Let alone to prevent from happening. The project needs some tools to help in QA improvements and better communication of (recently introduced) problems.

MonetDB Stories (ca 3 months, journalists)
Publications on MonetDB functionality targeted at a larger audiance are needed.

MonetDB promotional applications
MonetDB finds its way in many areas outside our direct view. Feedback and end-user experiences are therefore hard to get. Students working in close contact with industry who would like to port their existing applications, on a trial basis, to the MonetDB platform are invited to contact us.

PhD thesis topics

The MonetDB experimentation platform provides a sound basis to explore many areas of database kernel research. Its open-source and well-documented code base permits exploration of novel data structures, algorithms and optimization techniques geared at exploring the newest hardware developments. The MonetDB product family road-map provides ample opportunities to work at the frontier of database research and be able to demonstrate your innovation.

The algebraic interface over a binary storage model has been proven pivotal in developing alternative (query) languages on top of MonetDB. Areas of particular interest are query language support for (multimedia) information retrieval, which call for a completely different (probabilistic) evaluation model. Likewise, the system could be extended in the area of (time) series, P2P-systems, multi-core systems and streaming applications.