As a first postulate, there is no uniform data management solution for science. The use cases under development demonstrate the variety of data management functionality required. The demos are all constructed with MonetDB. They are examplars driving the development of the SciQL query language.
Another candidate to consider for scientific databases is PostgreSQL when there is a need for well-documented row-store database technology. Both MonetDB and Postgresql provide a rich programming interface and  come with tools for resource management. A complementary approach is taken by SciDB, a US-based project with similar objectives.
Data-intensive research requires an open and accesible architecture to convey the domain specific functionality at the right point in the software stack. A hindrance quickly turns into a processing nightmare, bearing the DBMS effectively useless for the task at hand.

… An open source solution is a sine-qua-non...