An organization facing big data pressures is migrating from VSAM to MongoDB. Although VSAM is 1970s technology, it stores data hierarchically which is very similar in structure to NoSQL document-based databases such as MongoDB. Therefore the data migration from VSAM to MongoDB could be straightforward. However, there is a debate going on whether a fully normalized relational logical data model should be built prior to migrating to MongoDB. Normalizing leads to well-understood structures and enables integration with existing data models. Some feel though that the effort to normalize will not be worth the value, because the current lack of metadata in the VSAM structures will lead to many unanswered questions that will only get answered once users can access the data in MongoDB. Also, there is a current lack of tools available for maintaining the lineage between the relational logical and the NoSQL physical, so the logical data model built may quickly become out of sync with its physical. What are your thoughts?
From VSAM to MongoDB