Sharing information among heterogeneous manufacturing databases through synchronization can lead to significant improvements in organizational productivity. Synchronization is hampered by the system and data complexities, proprietary nature of local database structures, and representation of similar information in dissimilar formats, and levels of granularity and abstraction. Using data translators for synchronization results in pairwise solutions and a loss of associativity between the updated and locally generated information. A methodology to synchronize multiple heterogeneous manufacturing databases is developed based upon the Integrated Manufacturing and Design Environment (IMDE) architecture. Multiple PDES/STEP databases are merged into a unified consistent representation that is used to update the local database of a given application. A neutral data representation for general manufacturing information is developed for use in the merge process.
The operations, data, and workflow of several organizations and data structures of typical software systems in various manufacturing functional domains were analyzed. A Unified Data Meta Model (UDMM) EXPRESS schema was developed to represent general manufacturing information in a neutral form. Software systems instantiate a relevant subset of the UDMM and map their local data to the UDMM format using intelligent translators. Procedures and algorithms to synchronize these multiple neutral representations and update the local software databases were developed. Two steps were necessary for synchronization: 1) external merge, where multiple representations were merged into a unified consistent representation, and 2) internal merge where the unified representation was used to update the local database.
The UDMM is an extensible, virtual schema that contains the union of the intersection of the information in the various manufacturing systems. It is non-persistent and UDMM instantiations are created upon synchronization requests. Variations in levels of detail, granularity, and abstraction are supported. The merge algorithms are data driven and a set of rules, defined for each software system, are used to synchronize information. The rules provide flexibility in the synchronization and allow different merge behavior at various stages in the product lifecycle. The merge algorithms are generalized and may be used to synchronize data between general PDES/STEP databases.