Automatically Denormalizing Document Relationships
Chief Software Engineer
Native XML databases provide no exception to the problem that data may not be easily contained by any single data storage idiom. Many-to-many relationships, in particular, present a unique problem for documents, as strategies for joining across documents are a potential minefield of software maintenance and performance problems. Automatic denormalization shifts the responsibilty for managing relationships to write-time, making an explicit trade-off for simplicity and speed at runtime. This paper discusses existing strategies for managing relationships across documents and explores design patterns and use cases for performing automatic denormalization and their trade-offs.