Researchers Peter Boncz, Stefan Manegold, and Martin Kersten of the Centrum Wiskunde & Informatica (CWI) in Amsterdam won the prestigious VLDB 10-year Best Paper Award on 27 August 2009. They received it for their article ‘Database Architecture Optimized for the New Bottleneck: Memory Access’. The award was presented at the 35th International Conference on Very Large Data Bases (VLDB) in Lyon, France. The VLDB conference is one of the most important international forums for database researchers, vendors, application developers and users. The winning article is about exploiting the capabilities of modern hardware using new database technology.
Relational database systems were designed in the late '70s and early '80s of the last century. At the time database performance was dominated by the time needed to read and write data on a hard disk of a computer. The central processing unit (CPU) was slow and not much data at a time could be loaded into the small main memory. Over time the hardware improved: CPU speed and memory size doubled roughly every 18 months. However, the design of database systems was not adapted. The database group of the CWI described in their winning article the first ideas about how database code and data structures should be adapted to the changed hardware. This was the beginning of a new approach to the discipline: hardware-aware database design. It allows for the full power of modern hardware to be utilized.
The award was also granted to the CWI researchers for their pioneering research in column-oriented database technology. Their open-source database system MonetDB makes modern column-oriented and hardware-aware database technology available to researchers, users and database developers. The VLDB 10-year Best Paper Award is annually granted for the best article that was presented at the VLDB conference ten years ago, and that has had the most impact since then. Of 50 selected papers the one of the CWI researchers was chosen. Controlling the data explosion is one of the main research themes of CWI. This research is a good example.
