Back to overview

Efficient Data Management for Scientific Applications

English title Efficient Data Management for Scientific Applications
Applicant Ailamaki Anastasia
Number 117125
Funding scheme EURYI (European Young Investigators)
Research institution School of Computer and Communication Sciences EPFL
Institution of higher education EPF Lausanne - EPFL
Main discipline Information Technology
Start/End 01.04.2008 - 31.03.2013
Approved amount 1'787'163.00
Show all

Keywords (5)

Database Management Systems; Scientific Workflows; Query Processing; Indexing; Automatic Data Placement

Lay Summary (English)

Lay summary
Computer-based applications are nowadays not only part of everyone's life, but also the basis of much scientific research, ranging from medicine and biology to astronomy to mechanical engineering. Computer programs that serve these scientific disciplines typically manage enormous amounts of data, which are stored in the computer's disks and are manipulated by the application program. Astronomers, for example, while trying to discover the shape of the universe, conduct their research by asking questions to a vast database consisting of sky object observations from the past several years.Computer science should be offering adequate solutions through database management system technology: a 40-year-old scientific discipline that deals with answering questions on large amounts of data, efficiently storing it on the disk, and massaging it to create answers. Unfortunately, database systems were born and grown to serve banking requirements (such as bank transactions) and more recently data warehouses (answering questions such as stock trend discovery). Scientists, however, require functionality that is not by modern database technology leaving scientific applications severely constrained by the complexity of manipulating massive datasets.This project aims at advancing database technology in order to help scientific groups manage their data efficiently. We differentiate between two groups of scientific work with respect to their data management requirements: Observation-based sciences (such as astronomy) face challenging data organization problems, whereas the difficulty when managing simulation-based data (such as earthquake modeling) is the complexity of computations on sophisticated data structures. In this project we will first analyze and understand the impact of the needs of these applications on data management. Then, we will translate what we learn into techniques that improve the efficiency of the supporting data management functionality. Finally, we will develop mechanisms to automate database design and administration to alleviate the burden to the scientists and to minimize human error. Through our efforts we hope to both improve the performance of scientific work as well as to simplify domain-specific programming, thereby enabling higher-impact scientific research.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants