In recent years, numerous new approaches that rely on data-intensive methods have been developed for maritime assistance systems, leading to a compelling need for more elaborate verification and validation procedures. Modern testbeds that can meet these demands are often developed separately from the system itself and provided as generically usable services. However, the joint usage of such testbeds by multiple stakeholders from research and industry confronts them with various challenges in terms of data management: Data control and protection is required to preserve possible competitive advantages or comply with legal framework conditions. The resulting decentralization in data management complicates collaboration, especially in the joint processing and analysis of testbed data. In this paper, we present a decentralized software system, which can deal with these challenges by modelling interrelationships between the stakeholders in a data space, considering their various interests. With the help of a modular data management architecture, the organization of a testbed data basis, as well as the support of verification and validation processes and the evaluation of data streams is made possible. This is achieved with a workflow model for mapping complex and distributed data processing steps. We demonstrate the applicability of the system in an application scenario for the development of a maritime assistance system.