A variational database management system

Abstract

Many problems require working with data that varies in its structure and content. Current approaches, such as schema evolution or data integration tools, are highly tailored to specific kinds of variation in databases. While these approaches work well in their roles, they do not address all kinds of variation and do address the interaction of different kinds of variation in databases. In this paper, we define a framework for capturing variation as a generic and orthogonal con- cern in relational databases. We define variational schemas, variational databases, and variational queries for capturing variation in the structure, content, and information needs of relational databases, respectively. We define a type system that ensures variational queries are consistent with respect to a variational schema. Finally, we design and implement a variational database management system as an abstraction layer over a traditional relational database management system. Using previously developed use cases, we show the feasibility of our framework and demonstrate the performance of different approaches used in our system