Difficulty: Medium
Correct Answer: A Bus Schema is a data warehouse design that uses a set of conformed dimensions and standardized fact definitions so that multiple data marts can plug into a common enterprise bus.
Explanation:
Introduction / Context:
In dimensional modeling and data warehouse architecture, the term Bus Schema comes from the approach popularised by Ralph Kimball. It is used to describe an enterprise level design in which different data marts share a common set of conformed dimensions and standardized facts. Interviewers ask about Bus Schemas to test understanding of how separate subject area marts can integrate into a unified warehouse that supports cross functional reporting.
Given Data / Assumptions:
Concept / Approach:
A Bus Schema is an architectural blueprint that identifies key business processes and the conformed dimensions that link them. It acts like a bus because each new data mart can plug into this shared set of dimensions and fact definitions. For example, both a sales mart and a returns mart might share the same customer, product, time and geography dimensions. By enforcing conformance, metrics become comparable and consistent across reports. The Bus Schema is often documented as a matrix showing processes as rows and conformed dimensions as columns. It guides the design of individual star schemas while preserving an overall integrated structure.
Step-by-Step Solution:
Step 1: Identify the major business processes to be modelled, such as order entry, shipment, invoicing and inventory movements.
Step 2: Identify candidate dimensions that apply across multiple processes, such as customer, product, time, region and channel.
Step 3: Define these dimensions in a conformed way, meaning that their keys, attributes and hierarchies are consistent and reusable across different fact tables.
Step 4: Lay out a Bus Matrix that maps each business process to the dimensions it uses, forming the conceptual Bus Schema.
Step 5: Implement individual star schemas for each process, using the conformed dimensions from the bus so that data marts plug into a unified enterprise warehouse.
Verification / Alternative check:
Imagine a company that first builds a sales data mart and later builds a service data mart. Without a Bus Schema, the sales team might define customers one way and the service team another way, leading to inconsistent reports. With a Bus Schema, the enterprise defines a single conformed customer dimension with shared keys and attributes. Both the sales fact table and the service fact table use this same dimension. As a result, analysts can easily create reports that combine sales and service metrics by customer. This integration benefit confirms that a Bus Schema is more than a single star schema; it is an enterprise bus of conformed dimensions and standardized facts, as described in option A.
Why Other Options Are Wrong:
Option B describes a fully normalised relational design with no dimensions, which is closer to an enterprise data model or third normal form warehouse, not a Bus Schema in dimensional modeling. Option C limits the concept to ETL job monitoring, which is not the purpose of a Bus Schema. Option D refers to a network topology metaphor for table connections, which is not standard in data warehousing. Option E treats Bus Schema as just another name for a star schema, ignoring the enterprise wide conformed dimension concept. Only option A correctly defines a Bus Schema as a shared set of conformed dimensions and fact definitions that multiple data marts can plug into.
Common Pitfalls:
A frequent mistake is to design each data mart independently without conformed dimensions, resulting in silos that cannot be easily integrated later. Another pitfall is to think that the Bus Schema requires building everything at once. In fact, the approach supports incremental delivery as long as each new mart follows the bus standards. In interviews, emphasise that the Bus Schema is a planning and integration tool: it ensures that as the data warehouse grows, the pieces fit together through shared dimensions and definitions, enabling consistent enterprise reporting.
Final Answer:
A Bus Schema is a data warehouse design that uses a set of conformed dimensions and standardized fact definitions so that multiple data marts can plug into a common enterprise bus.
Discussion & Comments