The Technology Research Of PDM Data Migration To NoSQL

At present,all kinds of datas involving PDM are stored in a traditional relational database,and presented to the user in the form of product structure tree in the manufacturing enterprise.At the process of dealing with all kinds of PDM data for storage, in order to conform to the characteristics of the relational database,we put a number of structured PDM data into a relational model. However, as the MBE/MBD technology widely implemented in manufacturing companies, the scale of the PDM data sharply increases, the processing strategies in dealing efficient querying and storage to the mass data, high reading and writing in concurrency, high scalability problems can't be dealt with very well. Therefore, looking for a kind of method which can migrate PDM data stored in relational database currently to the storage form that can effectively deal with the problem mentioned in preceding part of the text,is of great scientific research value.First of all, according to the current problems which are persecuting manufacturing companies in PDM data management, and the advantages of NoSQL in terms of unstructured data storage and fetching huge amounts of data, I put forward a kind of design pattern which migrated PDM data stored in relational database into NoSQL to solve the difficulties faced by manufacturing enterprises.Secondly, combining with the characteristics of the PDM data type, this paper analyzes the key need related to enterprise in terms of PDM data storage. Compared with the paradigm theory of relational database, this paper designed patterns that is specifically for PDM data migration, and the paper provide the specific process of the design patterns, and has completed the whole process of PDM data migration.Finally, we compared respectively PDM data stored in the relational database with PMD data processed by the migration method which MongoDB comes with, as well as PDM data processed in the migration patterns designed in this paper in terms of data storage and querying time. The experimental results show that the migration pattern in this paper is superior to the previous two in the query efficiency, and the effection is more obvious in thecase of large numbers of records.