Model-based Development of a Federated Database Infrastructure to support the Usability of Cross-Domain Information Systems

Abstract

The Centre for the Study of Manuscript Cultures (CSMC) at the Universität Hamburg hosts a steadily growing number of autonomously developed databases. These databases often run in heterogeneous hardware and software environments and use different, project-specific data models. These conditions lead to a high degree of autonomy but to an increasing heterogeneity. The heterogeneity then leads to the problem that these databases cannot be easily merged into a federated database system (FDBS). In general, a FDBS integrates multiple autonomous database systems into a single FDBS. The challenge is now to merge multiple heterogeneous systems into one FDBS to provide successful federated searches. Here, the term successful means that the results of a search query are correctly and completely answered by all integrated domain-specific database systems. In addition to delivering correct and complete results from these queries, semantically and syntactically correct mappings of various database schemes must be ensured. We argue that at the faculty of the Humanities, a cross-domain use of autonomously developed databases offers added value in terms of data analysis and evaluation. Hence, the challenge that exists is to develop a suitable federated database infrastructure that supports the development of autonomous information systems, while heterogeneity does not increase. To meet this challenge, in IT projects in the defense sector and in the public administration of the federal government, the V-model is a mandatory procedural standard to develop IT systems systematically. When developing a cross-domain information system, special attention should be paid to the development of variants, i.e. the project-specific information systems, so that the extension of the federated infrastructure at the faculty is guaranteed and can be maintained in the long term. In the area of Model-based Systems Engineering (MBSE) variant modeling is a very current topic and no standard exists. However, there are some approaches that can be used, such as the method Variant Modeling with SysML (VAMOS). VAMOS is characterized by its ease of use to model variants and can be combined very well with other methods. In this paper we present a model-based approach to develop a federated database infrastructure for supporting the usability of a cross-domain information system. For the development of a federated database infrastructure we present how the V-model can be combined with the VAMOS method using the tool Cameo Systems Modeler for modeling and simulating in general, the broker-based SysML Toolbox for simulating database interactions, and the database management system Heurist for creating a cross-domain information system.