“In which a small country can be great” – that is a rough translation of a Dutch expression that is sometimes used as a motto for the country to lift itself up. The Netherlands sometimes suffers from the so-called Calimero syndrome, a mental attitude derived from an Italian cartoon series in which a tiny black chick called Calimero complains in a high-pitched voice: “It’s not fair! Because they are big and I am small.”
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
But as a relatively small country, The Netherlands has contributed big things to the worldwide computer industry. IT experts with historical knowledge will have heard of computer scientists Edsger Dijkstra and Andrew Tanenbaum. Both these Dutchmen have received several prestigious computer science awards, including the US-based Association for Computing Machinery (ACM) Fellow Award.
Now another Dutchman is joining this illustrious company: Martin Kersten, for his many years of work on database design. He has not just worked on database theories, but has produced interesting possibilities that might be achievable in the future. The Netherlands’ contribution to databases is very much practical and in the here and now. There is actual software and real-world products, a functioning company and a community based around open source software – MonetDB.
Amsterdam-based Kersten is level-headed about his ACM award, joining the very small number of Dutchmen to become ACM Fellows over the years. In fact, the number of non-Americans on that list is also quite small. Of the 50 Fellowships awarded last year, only seven were for computer scientists outside the US, and Kersten is one of those.
“It is good for our international prominence,” says Kersten. A professor at Centrum Wiskunde & Informatica (CWI), The Netherlands’ national research institute for mathematics and computer science, he says the award means international recognition for CWI and the country as a whole.
Kersten points out that CWI, where he has worked on database technology since 1985, has an impressive track record. “In the 1950s, the first computer was built here,” he says. “And the Python programming language was made here, by Guido van Rossum.” This particular CWI invention is widely known and used in the IT world.
There is an important distinction between CWI and other computer science research centres. The Dutch institute is not focused on scientific research for research’s sake, but aims to produce practical applications of technology. “As well as research, we look at the long term with products going to market in mind,” says Kersten.
The CWI is an executive laboratory and has been called “a university without students”.
Kersten is now a standard-bearer for Dutch contributions to IT. His 2016 ACM Fellow Award, officially presented in June this year, was for a wide range of database design work. A specific reason given was “his work on the design and implementation of column store and main memory database systems”.
In-memory databases are now commonly used in the IT world, as is column storage of information in databases. Such technology is required for modern applications employing large amounts of data in various ways.
Kersten says: “How would you describe a database to your mother? It’s like making vegetable soup.” But the professor is not dumbing it down – his comparison makes sense: there are ingredients, a recipe for those ingredients, and only at the end is the flavour achieved. “It’s all about the overall picture,” he says.
Database design is similar in that it is strongly architecturally oriented, says Kersten. The traditional structure of databases comprises pages with tables, which works quite well if you are interested in one particular line within a specific table of a database, he says. “Like, for example, a person’s bank account.” Known as on-line transaction processing (OLTP), this is used extensively in the financial world.
Big data, deep learning
A different approach, for use with multiple sets of data, is on-line analytic processing (OLAP). This is a column-based structure for databases that speeds up queries into multiple datasets. OLAP has proved its worth in applications such as business intelligence, data mining and big data.
MonetDB is a real-world implementation of this approach, and the result of Kersten’s research at CWI. “In a traditional database, the DBA needs to make or choose indexes of the data,” he says. “However, that is not necessary for column-based databases.” This enables far speedier processing of certain queries for specific applications, he adds.
A good use example is deep learning, previously known as data mining. Back in 1995, Kersten was working on data mining, which resulted in CWI spin-off Data Distilleries. This Dutch company used early implementations of MonetDB to produce analytical software for customer relationship management (CRM). The company was acquired by SPSS, which itself was then acquired, putting the technology into the hands of IBM.
Recognition and chest-beating
But this US commercialisation was no deterrent to continued Dutch development of database design. Following on from Data Distilleries came another CWI spin-off, VectorWise, and more recently, in 2013, MonetDB Solutions was created. The aim of that company is to promote usage of MonetDB in the market, says Kersten. “For example, to develop functionality that is perhaps not so interesting scientifically, but is of interest for business use,” he adds.
Meanwhile, the research work continues – as does the gradual recognition of Dutch database design. Kersten says The Netherlands should be more proactive in pushing the diverse range of IT research, development and innovation that is taking place in the country. “We don’t really beat our chests,” he says. “The Americans are much better at that. We could really follow their example.”