A polyglot persistence approach is being used by storing master data and key relationships in Neo4j (wide data) and integration with the microservice framework linking source data (deep data).
Three key points:
- Product DNA via Master Data
Through the extraction and storage of master data and key relationships the DNA of our products at Lockheed Martin Space is revealed. The capability to find correlations between business entities that we assumed to be unrelated or orthogonal emerges. This provides insight needed to identify more agile and efficient processes throughout the product lifecycle.
- Polyglot Persistence Model
Only the master data that forms the digital threads across the product lifecycle is uploaded to the graph using streaming technologies. We describe this as our wide data. The master graph is integrated with the microservice catalog linking the master data to all other data via microservices. We call that data our deep data. All supporting data is left in either the transaction or data warehouse systems. Additionally, an access control framework independent of the transaction systems had to be developed to support policies for blended data sets.
Graph technology is not your grandma’s data architecture. By getting training and services from experts a successful architecture can by developed. Years of experience modeling in traditional RDBMS can be a hindrance.