Informationsfusion und Integritt: GridErweiterungen zum Datenmanagement

1
Informationsfusion und -IntegritätGrid-Erweiteru
ngen zum Datenmanagement
Project Leader Prof. Dr.-Ing. Erhard
Rahm Institut für Informatik, Universität Leipzig

• Project Partners
• Infrastructure
• TU Dresden
• Universität Leipzig
• TU München
• Application
• AstroGrid-D
• MediGrid
• TextGrid
• Forschungszentrum Rossendorf

2
Main Goals
Services for

• Data Fusion
• semantical correct Combination of Data Sources
• Support of Analysis over distributed Data
• Ontology-based Mapping of Data Sources
• Data Integrity
• Data Lineage
• based on Grid Security Infrastructure (GSI)
• Dynamic Data Distribution (see following slides)
• Optimization of data intensive Queries
• effective and efficient Usage of decentral Main
  Memory Ressources

Augmenting Data Management in DGI
3
HiSbase Main Memory P2P Database System
4
HiSbase The Challenge

• Histogram-based peer-to-peer main memory database
  for locality-aware data processing
• Example distributed archives (astrophysics)
• Correlation of different catalogs
• Skewed data distribution
• Region-based queries
• Right ascension (longitude)
• Declination (latitude)

5
(No Transcript)
6
Distribute by Region not by Archive!

• Highly distributed information management
• Distributed Hashtable peer-to-peer architecture
• High performance query processing
• Main memory database
• Semantic clustering, spatial locality
• Equi-depth histograms

7
(No Transcript)