Hi everyone!

We are wrapping up our introduction to the MDM series with this post delving into the core of MDM—the Model. MDM is a model-driven implementation, where the model takes center stage, driving the generation of structures and orchestrating crucial processes. Let’s dive in
 

Unlocking the MDM Model Magic

The MDM Model is the maestro orchestrating the symphony of Ataccama MDM's capabilities. Here's a glimpse of how it weaves its magic:

• Automated Database Structures: Say goodbye to manual database creation. Ataccama MDM automatically generates database structures based on your model.

• MDM Workflow Wizardry: Cleansing, matching, and merging—the three crucial phases—are seamlessly handled by the MDM Workflow, automating the orchestration of these intricate processes.

• Native MDM Services Creation: The model isn't just a blueprint; it's a generator. Native MDM services spring to life based on your model, simplifying service creation and eliminating unnecessary complexities.

• Convention Over Configuration: Embrace simplicity with the model's convention over configuration approach, streamlining processes by utilizing model information.
   

  

  
   

A Visual Tour of the MDM Model

To harness the modeling prowess of Ataccama MDM, developers wield a robust modeling interface. The MDM model project, as depicted below, showcases the intricate relationship with various elements within the MDM architecture.

Ataccama MDM embraces the power of data models created visually through an in-built case tool. This tool empowers developers to define comprehensive data models, including master data, data sources, data consumers, and general input/output interfaces. The model's metadata resides within MDM, offering flexibility and independence from source system structures.

Models in Ataccama MDM transcend the boundaries of original data, allowing developers to create entities and structures independently of the source system. This abstraction leads to the creation of a common canonical model, facilitating a unified understanding.

As an XML-based metadata layer, MDM effortlessly imports models from external case tools using standard metadata exchange formats (XMI, XSD) and native formats (Power Designer, ERwin). You can also import models directly from a database, ensuring seamless integration and adaptability.
 

Instance Layer: Where Source Meets Cleansed Form

In the realm of MDM, the instance layer takes center stage, serving multiple purposes:

• Accommodating Source Attributes: It creates a structural haven to house diverse attributes from various sources, ensuring flexibility and adaptability.

• Defining Target Structure: The instance layer shapes the target structure for the cleansing phase, acting as a canvas for the transformation of source records into their cleansed forms.

• Metadata Hub for Matching: Holding the metadata related to the matching process, the instance layer plays a pivotal role in the journey from source to cleansed records.

• Relationships Unveiled: Importing relationships from the source data, this layer unveils the intricate connections between different records, adding depth to the understanding of data entities.
   

Master Layer: After the MDM Phases Dance

Once the MDM phases—cleansing, matching, and merging—have performed their dance, the master layer takes the spotlight. It's a composition of models representing entities, relationships, and attributes.

In the realm of the master layer:

• Unification of Source Data: Post-MDM phases, the master layer unifies source data into a common master structure, bringing harmony to the diverse datasets.

• Golden Records and Deduplication: Defining which instance records correspond to each master record, it marks entities as either golden records or deduplicated instances.

• Relationships Rule: Enforcing relationships between master entities automatically, the master layer becomes the base for Master Data Admin (MDA) to interact with master data.
   

Physical Layer: Bridging the Logical and Tangible

The physical layer of the MDM repository translates the conceptual into the tangible, with unique tables for each entity at the instance and master levels. Here's how it unfolds:

Instance Layer:

• L1: Stores source data, aligning with the canonical structure devised to accommodate all heterogeneous structures from connected systems.

• L2: Homes cleansed data, including standardized attributes and metadata related to the cleansing process.

• L3: A virtual layer holding metadata associated with the matching process for selected entities.

• L3.5: An aggregation layer allowing record aggregation based on a chosen attribute.

Master Layer:

• L4: Holds master data, including merged results from matching, relationship keys, and master data for both mastered and deduplicated entities.

• L4.5: Stores validation scores and explanation codes for all master data, providing insights into data quality grades and attribute validation severity.
   

  

  
   

Configuration: Setting the Stage for Success

Beyond data models, several essential configurations contribute to a successful MDM project:

• Connected Systems Definitions: Define the systems that MDM interacts with.

• Cleansing, Validation, Match, and Merge Rules: Fine-tune the rules that shape your MDM processes.

• Aggregation Logic: Configure logic for aggregating records and propagating values to master layers.

• Service Definitions: Set up services provided to connected systems, such as exports, event handlers, and publishers.

• Reprocessing and Auditing: Configure additional features, authorizations, and audit levels to tailor MDM to your specific needs.
   

Transformation (MDM Phases): The Symphony Unfolds

The MDM process unfolds in three distinct phases, each playing a crucial role in shaping and refining your data:

Cleansing:

Leverage the powerful DQC engine to ensure data quality. MDM provides a range of pre-built operations and functions, allowing for flexibility and high-performance data cleansing.

Matching:

Harness a state-of-the-art, high-performance rule-based matching engine that supports multiple strategies, scoring-based rule matching, and ID stability for key business processes.

Aggregating:

Optional yet powerful, the Aggregation layer enables the calculation of aggregate values on records grouped by selected attributes. It adds a layer of depth to your data, allowing for versatile insights.

Merging:

MDM's high-performance merging operation, combined with metadata from matching, offers flexibility in selecting data for the golden record. Rule-based merging ensures precise control over the merging process.

Validating:

Translating explanation codes to validation messages, MDM calculates data quality grades and attribute error severity grades. These insights are presented in the MDA user interface, offering a comprehensive view of data quality.

Online Service Interfaces:

Native MDM services, automatically generated based on your model, provide SOAP or XML services for accessing hub data. The flexibility to configure services ensures that you have the tools you need at your fingertips.
 

With the model as your guide, you navigate through intricate layers, ensuring harmonious data integration and mastery. Stay tuned for more