Benchmarking Tool for Comparing Semantic Dataset Storage Solutions

OData support
Simon Gábor
Department of Automation and Applied Informatics

Nowadays as the different NoSQL based database managers are living their golden age, new solutions appear more frequently using completely different concepts. This kind of versatility does not avoid the graph databases as well. With the appearance and growing popularity of the multimodal database managers the graph-based solutions are beginning to appear in such classic fields as the relation database managers (for example one of the biggest new feature of the Microsoft SQL Server 2017 is the support of the graph-like queries).

At the same time, as the globalization is slowly reaching its totality, the quantity and the importance of information is growing in an unprecedented way, so it's quite natural that the question of validity or reliability arises, and how we represent it in computers, which lead us straight towards the general problem of reification (how we store information about information).

Regarding the usability of a system today, its performance is almost as important as the realized functionality, so it's critical to use this kind of information in the most efficient way possible. The first step towards it is selecting the proper representation model and physical storage mode.

In my paper, I compare the classic modeling solutions to the problem of reification on the available database managers with different concepts which offer graph-based data processing. My focus is on the performance of the different models on different storages, that will help to answer what database manager and data model should be used to gain the best performance on a reified dataset with predefined characteristic used by queries belonging to a predefined kind of query group.


Please sign in to download the files of this thesis.