Modeling Data Vault for Scalable Architectures

Modeling Data Vault for Scalable Architectures

In this lesson, you will learn how to build highly agile and audit-ready data structures using the Data Vault methodology to handle complex, evolving business environments.

When you move beyond traditional dimensional models, you often find that rigid schemas struggle to accommodate rapid source system changes. Data Vault solves this by breaking your data into three distinct components: Hubs, Links, and Satellites.

The Core Components

1. Hubs (Business Keys): A Hub represents a core business concept, such as a Customer or Product. It stores only the natural business key (e.g., customer_id) and a surrogate key. Hubs provide the structural backbone that never changes, even if you add new data sources.
2. Links (Relationships): A Link represents the relationship between two or more Hubs, such as a Sale connecting a Customer to a Product. By separating relationships into their own tables, you can add new types of associations without restructuring your core entities.
3. Satellites (Descriptive Data): Satellites store the "context" or attributes of your data, like a customer's address or email. Because they are linked to a Hub or a Link, you can track changes over time (auditability) by adding rows rather than updating records.

Illustrative Example

Imagine you need to track a Customer buying a Product. Instead of one massive table, you distribute the data:

-- Hub: The core business key
CREATE TABLE hub_customer (hk_customer_id, customer_number, load_date);

-- Link: The relationship between customer and product
CREATE TABLE link_sale (hk_sale_id, hk_customer_id, hk_product_id, load_date);

-- Satellite: The changing descriptive attributes
CREATE TABLE sat_customer_details (hk_customer_id, address, phone, load_date);

By decoupling your keys, relationships, and attributes, you create a "LEGO-like" architecture. If a source system changes how it stores customer details, you simply create a new Satellite. Your Hubs and Links remain untouched, ensuring your pipeline remains resilient and fully audit-ready. This modularity is what makes Data Vault exceptionally scalable for modern, distributed data environments.

Key Takeaway: By separating business keys (Hubs), relationships (Links), and descriptive context (Satellites), you create a modular architecture that allows you to add new data sources without disrupting existing analytical structures.