Heinrich-Heine-Universität Düsseldorf

Vita

Bachelor degree at the research group for Database and Information Systems supervised by Prof. Dr. Conrad. It is about developing a system for querying RDF-data that is stored in relational databases, called RDQuery. Queries were formulated in SPARQL and translated into SQL. Results are translated back into RDF-statements. Title of the thesis: “Translation of RDQL-Queries into SQL” [Paper]

Master degree at the research group for Software Techniques and Programming Languages supervised by Prof. Dr. Leuschel. I wrote a [ProB]-plugin for the symbolic model checker [SMV] and an Eclipse-Plugin for the NuSMV syntax. Title of the thesis: “Validating SMV Specifications in ProB”.

Current Research

Relational databases (RDB) are often the first choice, when it comes to the storage of application data. With the different structural properties of a domain model, e.g. tabular or graph-based, a RDB is not always the ”best” choice, regarding optimal storage and query performance. With the development of new storage types, e.g. Triple Stores, NoSQL-Databases like Key-Value stores and Graph-databases, etc., which produce other efficient schema designs, the choice for the most suitable storage type for a domain model is far from trivial.

In my research, we evaluate different storage types and focus on the aspect of query performance. We choose domain models from the social sciences and data from a real life system, together with typical use case queries for this domain model, which are widely accepted by the domain experts. Evaluations show that a RDB is not always the best choice with respect to query performance of some queries and that a closer look at the schema might influence the choice for the most suitable storage type.