data vault - ghdrako/doc_snipets GitHub Wiki

Dan Linstedt’s Data Vault modeling is a methodology for long-term historical storage of data coming in from multiple operational systems to a unique database. It’s also a way of looking at historical data that addresses issues such as auditing, data tracing, and loading speed. Its main components are hubs, links, and satellites.

Hubs are dedicated to storing unique business keys along with their descriptions. Each hub represents a unique business concept or object, such as a customer or product. Links, as the name suggests, are about connections. They store the associations or relationships between business keys, effectively capturing the many-to-many relationships between business objects. Satellites store all contextual or descriptive data. They hold the attributes related to the business keys (in hubs) or their relationships (in links). Importantly, satellites also capture the time-variant nature of this descriptive data, thereby providing a temporal perspective.