Multi-Active Satellites vs. Dependent Child Links
In the realm of Data Vault modeling, practitioners often encounter scenarios where multiple descriptive entries are valid simultaneously for a single business entity. Two primary modeling techniques address this complexity: Multi-Active Satellites (MAS) and Dependent Child Links. Understanding the distinctions between these approaches is crucial for designing efficient and accurate data warehouses.
In this article:
Understanding Multi-Active Satellites
A Multi-Active Satellite is designed to store multiple instances of descriptive information related to a parent key, all valid at the same point in time. The parent can be either a Hub or a Link. This structure is particularly useful when an entity can have several concurrent attributes.
For example, consider an insurance policy that offers various coverage details, each with different validity periods. Here, the policy (Hub) is associated with multiple coverages, each represented as a row in the Multi-Active Satellite, capturing the distinct validity periods and coverage amounts.
Defining Dependent Child Links
A Dependent Child Link is a Link entity that includes one or more additional key attributes. Together with the combination of business keys connected by the Link, these attributes uniquely identify incoming data records. This structure is also known as a degenerated link, peg-legged link, non-historized link, or transactional link.
For instance, in an invoicing system, an invoice (Hub) may have multiple line items. Each line item can be uniquely identified by combining the invoice identifier with a line item number, forming a Dependent Child Link.
Modeling Example: Order Line Items
When modeling order line items, there are two valid approaches:
- Create a non-historized, Dependent Child Link with a Non-historized Satellite containing the invoice footer details.
- Establish a non-historized, Dependent Child Link that includes both the key combination and the invoice footer details.
The choice between these methods depends on the specific requirements of the data model and the nature of the data being captured.
Modeling Example: Insurance Policies
Consider an insurance policy with different effective time windows—a scenario discussed in a previous session. In this case, attributes such as ValidFrom, ValidTo, and Amount are descriptive data attributes related to the business relationship between the Policy and Coverage.
The recommendation is to keep these attributes together in a Multi-Active Satellite on a standard Link between Policy and Coverage. This approach ensures that all relevant information is stored cohesively, allowing for efficient querying and analysis.
Choosing Between Multi-Active Satellites and Dependent Child Links
The decision to use a Multi-Active Satellite or a Dependent Child Link hinges on the specific business scenario:
- Multi-Active Satellites are ideal when multiple descriptive attributes of an entity are valid simultaneously, and these attributes need to be tracked over time. This structure allows for capturing the history of changes effectively.
- Dependent Child Links are suitable when there is a need to uniquely identify records through a combination of keys, especially in transactional contexts where multiple related records exist, such as invoice line items.
It’s essential to assess the nature of the data and the business requirements to determine the most appropriate modeling technique.
Conclusion
Both Multi-Active Satellites and Dependent Child Links offer valuable structures in Data Vault modeling, each catering to different scenarios involving multiple concurrent records. By understanding their definitions, applications, and differences, data modelers can make informed decisions to design robust and efficient data warehouses.
Watch the Video
Meet the Speaker
Trung Ta
Senior Consultant
Trung has been Senior BI Consultant since 2019. As a Certified Data Vault 2.0 Practitioner at Scalefree, his area of expertise includes Data Warehousing in a cloud environment, as well as Data Vault 2.0 modeling and implementation – especially, but not limited to, with Wherescape 3D/RED. He’s been working with industry leaders in the insurance and finance sector, advising them on building their own Data Vault 2.0 solution.
