Delta DB

Current Research Project

Summary / Zusammenfassung: In many data management applications, changes to the data are more important than the base data itself. One example are tumor databases, where the data does not comprise mutated DNA, but mutations of DNA. Other examples include the analysis of document evolution in version control systems (e.g. software source code) to investigate patterns and relationships, and implementation of synchronization strategies for distributed data (desktop, mobile device, cloud), where changes have to be merged or reconciled according to application-specific rules.

In this project, we investigate methods to manage changes of base data, by elevating the concept of change to a first-level type in the logical data model. Our model is based on the observation that in our applications of interest, both base data and changes between versions can be represented as ordered sequences of tuples. We propose a declarative query language for changes, to support application development in domains like the ones described above. The duality of changes and base data is exploited during query processing: The query processor may change the representation of data from materialized versions to deltas, and back, to improve performance.

Keywords / Suchbegriffe: delta database, version management, database systems, version, versions

Project Leadership and Contacts / Projektleitung und Kontakte

Dietrich Christopeit (Project Leader)

Funding Source(s) / Unterstützt durch: Universität Zürich (position pursuing an academic career)

In Collaboration with / In Zusammenarbeit mit

Prof. Dr. Carl-Christian Kanne, Database Research Group, University of Mannheim

Germany