Data Marketplaces are emerging digital platforms in data management. These platforms enable the infrastructure for data exchange, offering the opportunity for data providers to commercialize their datasets and save time for data consumers in identifying appropriate datasets. Conceptually, data marketplaces provide data exchange services as their core activity for the business community. However, in practice, organizations may be reluctant to exchange data through intermediaries as marketplaces instead of exchanging data bilaterally. The bilateral exchange of data gives more control to providers and consumers over the exchange process. For this purpose, a decentralized architecture could represent the peer-to-peer technical solution for the bilateral exchange. Therefore, what was previously called intermediaries, in the decentralized environment, data marketplaces are communication structures facilitating the commercial exchange of data between organizations. In this thesis, a specific focus on a decentralized architecture, namely blockchain technology, is considered for data marketplaces. The research follows the firm belief that both the technical and business specifications are essential in understanding the data marketplaces with blockchain architecture. Therefore, the decentralized architecture and the business models for data marketplaces, are researched with the overarching objective to investigate the impact of a decentralized architecture on the business models of data marketplaces. This research contributes to the literature by improving the general understanding of decentralized data marketplaces and by providing technical and business frameworks that can be utilized to investigate the application of blockchain technology in data trading platforms as well as analyze their business model characteristics.