Researchers of Centrum Wiskunde & Informatica (CWI) in Amsterdam have developed new methods to index huge databases. These space efficient indexes are lightweight to build and maintain and grow less than linear with the size of the data sets.
Researcher Eleftherios Sidirourgos will defend his thesis on this topic “Space efficient indexes for the Big Data Era” on Wednesday 14 May at the University of Amsterdam. The results of his research can be used for a variety of applications in science and e-commerce.
Researchers in both science and industry use the analysis of big data sets to make new discoveries, ranging from astronomical exploration to targeted advertising. These sets are continuously growing as new data keeps flowing in. In order to keep this manageable and searchable, there is a need for space-efficient indexes, data structures that are used to quickly locate data in a database.
In his thesis, Sidirourgos has developed a number of such indexes that address the requirements of a specific application. These include indexes for fast queries on scientific databases, for instance astronomical observation data, an index for use on large e-commerce sites that can put data on ‘hot’ users and products in fast memories, and ‘cold’ older data in slower memories.
This project is part of the Database Architectures research group at CWI, which aims to bridge the gap between database architectures and the need for scalable algorithms that are able to retrieve information from databases with a large number of elements. The results of this research are accumulated in MonetDB: open source database management software for very large databases.
Thesis: Space efficient indexes for the Big Data Era
By: Eleftherios Sidirourgos (CWI Database Architectures group)
Promoter: Prof. dr. M.L. Kersten (CWI/UvA)
Co-promoter: Prof. dr. P.A. Boncz (CWI/VU University Amsterdam)
Date: 14 May 2014, 13.00h
Place: Oude Lutherse Kerk, Singel 411, Amsterdam
