Smart manufacturing is characterized as transparent shop floor production, rapid and intelligent responses to dynamic changes, and a utilization of high-performance inter-cooperation networks. Smart manufacturing and a global appetite for personalized products have transitioned industry from mass production into the age of mass customization. Increased autonomy is slowly changing customer expectations as well, enabling customers to modify a product design not only during an order, but sometimes even long after placing an order. In this context, this paper fills a gap by presenting a data-centric infrastructure to enable interaction with a “global, virtual data space,” which overcomes the problems with traditional direct access methods such as interoperability and compatibility. Using a Cyber-Physical System (CPS), resource monitoring on the shopfloor as well as multiple parities beyond the enterprise boundary will be interconnected through this data-centric infrastructure. A semantic knowledge management system, which encompasses product lifecycle knowledge and manufacturing process ontology, is developed as the data schema in the data-centric infrastructure. In comparison to relational databases which are effective at handling paper forms and tabular structure, the flexible schema of graph databases enable these to handle dynamic and uncertain variables. These capabilities are deemed critical for a platform supporting real-time information exchange between customer, manufacturer and collaborators. One advantage of such a system allowing for real-time information exchange is that it enables last minute order changes by the customer, allowing for product design changes even after production has started on the order. The other advantage is that it allows manufacturing managers to monitor the productivity of customer-directed, dynamic manufacturing processes by utilizing Dynamic Value Stream Mapping (DVSM) methods.