Databases are getting bigger and bigger: an increase of data in the order of gigabytes or even terabytes is not unusual anymore. What to do if a database no longer fits on one computer? Could you make a database flexible and scaling independently with any expected growth? Fabian Groffen, PhD of the Centrum Wiskunde & Informatica (CWI) in Amsterdam, developed the Armada-model.
Fabian Groffen received his PhD on Wednesday, 10 June at the University of Amsterdam on his thesis: "Armada, an Evolving Database System" The model he developed can be used in the distributed storage of large scientific databases and data warehouses.
If large databases no longer fit on one computer, it is inevitable that more computers are needed. In this enlargement however, the whole database has to be installed again: effective cooperation between computers is not evident. Fabian Groffen investigated a self-managing database system that maintains itself. In his model, the computers act independently but they will help each other in case of an emergency, just as it happened in a 16th-century fleet (Armada). With this method there will be less pressure on the system administrators.
Mixed data
The ultimate goal of the research is to realize shared data that is flexibly expandable. Fabian Groffen investigated the flexibility and scalability of the model. He tested it on several prototypes and applied parts of his theory on the MonetDB database system, developed at CWI. The prototype proved that his ideas work. Managing the data explosion is a key factpr of the CWI research.
Listen to the dutch interview with Fabian Groffen on BNR Nieuws Radio
Fabian Groffen received his PhD on Wednesday, 10 June at the University of Amsterdam on his thesis: "Armada, an Evolving Database System" The model he developed can be used in the distributed storage of large scientific databases and data warehouses.
If large databases no longer fit on one computer, it is inevitable that more computers are needed. In this enlargement however, the whole database has to be installed again: effective cooperation between computers is not evident. Fabian Groffen investigated a self-managing database system that maintains itself. In his model, the computers act independently but they will help each other in case of an emergency, just as it happened in a 16th-century fleet (Armada). With this method there will be less pressure on the system administrators.
Mixed data
The ultimate goal of the research is to realize shared data that is flexibly expandable. Fabian Groffen investigated the flexibility and scalability of the model. He tested it on several prototypes and applied parts of his theory on the MonetDB database system, developed at CWI. The prototype proved that his ideas work. Managing the data explosion is a key factpr of the CWI research.
Listen to the dutch interview with Fabian Groffen on BNR Nieuws Radio
More information: http://homepages.cwi.nl/~fabian/.
Supervisor: prof. dr. M.L.Kersten (CWI en UvA)
Co-supervisor: dr. S. Manegold (CWI).
The research is funded by the Dutch BRICKS Bsik program and is carried out at the CWI under auspicies of SIKS, the Dutch Research School for Information and Knowledge Systems.
