About the Guided Stewardship operating model and data layers

Guided Stewardship offers the Guided Stewardship operating model, an out-of-the-box operating model to define the structure of the information in Collibra. The full operating model consists of 3 layers, representing the different structural data layers in typical organizations:

• The physical layer represents the actual storage-level structure of the data as they exist in the source systems. The physical layer consists of assets such as Schema, Table, and Column assets.
• The semantic layer, also referred to as the logical data layer, provides a business-centric view of data. The semantic layer consists of assets such as Data Model, Data Entity, and Data Attribute assets, which act as a bridge between the raw physical assets and the Knowledge Graph's governance assets, such as Business Term, KPI, and Data Category assets.
• The conceptual layer defines the enterprise blueprint for data, focusing on broad, system-independent Data Domains and Data Concepts. It provides a standardized way to understand data and serves as the foundation for organization-wide governance.

The Guided Stewardship operating model specifically includes the semantic layer and conceptual layer asset types, such as Data Attribute, Data Entity, Data Model, Data Concept, and Data Domain. The physical data layer asset types, Schema, Table, and Column, are available with Data Catalog.
Database and System assets are Technology assets that represent the highest level over physical data and logical data organization.

Tip The definition of all 3 layers is very useful for larger organizations that have data governance experience. For other organizations, defining the physical layer and the semantic layer can be a good starting point. If the full Guided Stewardship operating model is too complex for your organization, we advise to skip the conceptual layer because the semantic layer provides better out-of-the-box features and relation possibilities to the physical layer.

The following image shows all relevant asset types, per layer, and the relationships that bind them together in Collibra.

About the physical layer

The physical layer represents the actual data in an organization's systems. It consists of the following asset types: Database, Schema, Table, and Column.

• The Schema, Table, and Column asset types are available with Data Catalog.
  The Schema, Table and Column assets are almost never created manually. They are mostly automatically created via the Data Catalog registration processes.
• Although the Database asset type is a Technology asset, it is considered part of the physical layer.

About the semantic layer

A semantic layer makes it easier for business users to understand the meaning and context of data. The main building blocks of the semantic layer, Data Entity and Data Attributes assets, act as a bridge between physical data assets, such as databases or files, and governance assets in the Collibra Knowledge Graph, including business terms, KPIs, and data categories.

• The Data Entity and Data Attribute asset types are available with Guided Stewardship.
• Although the System asset type is a Technology asset, it is considered part of the semantic layer. Including a System asset is not mandatory in the Guided Stewardship operating model.

The semantic layer is a tree-like structure, starting with high-level System and Data Model assets, and branching out with implementation-specific Data Entity and Data Attribute assets.

{
    "nodes": [
        {
            "id": "Column",
            "type": {
                "id": "00000000-0000-0000-0000-000000031008"
            },
            "editorSettings": {
                "nodePropsExpanded": false
            },
            "layoutRegion": "flow"
        },
        {
            "id": "Data Attribute",
            "type": {
                "id": "00000000-0000-0000-0000-000000031005"
            }
        },
        {
            "id": "Table",
            "type": {
                "id": "00000000-0000-0000-0000-000000031007"
            },
            "display": "collapsed",
            "layoutRegion": "context"
        },
        {
            "id": "Schema",
            "type": {
                "id": "00000000-0000-0000-0001-000400000002"
            },
            "layoutRegion": "context"
        },
        {
            "id": "Data Entity",
            "type": {
                "id": "00000000-0000-0000-0000-000000031004"
            },
            "display": "collapsed",
            "layoutRegion": "context"
        },
        {
            "id": "Data Model",
            "type": {
                "id": "00000000-0000-0000-0000-000000031003"
            },
            "layoutRegion": "context"
        },
        {
            "layoutRegion": "context",
            "id": "Database",
            "type": {
                "id": "00000000-0000-0000-0000-000000031006"
            }
        },
        {
            "id": "Data Category",
            "type": {
                "id": "00000000-0000-0000-0000-000000031109"
            }
        },
        {
            "id": "KPI",
            "type": {
                "id": "00000000-0000-0000-0000-000000011002"
            }
        },
        {
            "id": "Business Term",
            "type": {
                "id": "00000000-0000-0000-0000-000000011001"
            }
        }
    ],
    "edges": [
        {
            "from": "Data Attribute",
            "to": "Column",
            "label": "",
            "style": "arrow",
            "type": {
                "id": "00000000-0000-0000-0000-000000007094"
            },
            "roleDirection": true
        },
        {
            "from": "Column",
            "to": "Table",
            "label": "",
            "style": "boxed",
            "type": {
                "id": "00000000-0000-0000-0000-000000007042"
            },
            "roleDirection": true
        },
        {
            "from": "Table",
            "to": "Schema",
            "label": "",
            "style": "boxed",
            "type": {
                "id": "00000000-0000-0000-0000-000000007043"
            },
            "roleDirection": false
        },
        {
            "from": "Data Attribute",
            "to": "Data Entity",
            "label": "",
            "style": "boxed",
            "type": {
                "id": "00000000-0000-0000-0000-000000007047"
            },
            "roleDirection": false
        },
        {
            "from": "Data Entity",
            "to": "Data Model",
            "label": "",
            "style": "boxed",
            "type": {
                "id": "00000000-0000-0000-0000-000000007046"
            },
            "roleDirection": true
        },
        {
            "from": "Schema",
            "to": "Database",
            "label": "",
            "style": "boxed",
            "type": {
                "id": "00000000-0000-0000-0000-000000007024"
            },
            "roleDirection": false
        },
        {
            "from": "Business Term",
            "to": "Data Attribute",
            "label": "",
            "style": "arrow",
            "type": {
                "id": "00000000-0000-0000-0000-000000007038"
            },
            "roleDirection": true
        },
        {
            "from": "Data Category",
            "to": "Data Attribute",
            "label": "",
            "style": "arrow",
            "type": {
                "id": "c0e00000-0000-0000-0000-000000007315"
            },
            "roleDirection": true
        },
        {
            "from": "KPI",
            "to": "Data Attribute",
            "label": "",
            "style": "arrow",
            "type": {
                "id": "00000000-0000-0000-0000-000000007200"
            },
            "roleDirection": true
        }
    ],
    "showOverview": false,
    "enableFilters": true,
    "showLabels": true,
    "showFields": true,
    "showLegend": true,
    "showPreview": true,
    "visitStrategy": "directed",
    "layout": "FlowContext",
    "maxNodeLabelLength": 50,
    "maxEdgeLabelLength": 30,
    "layoutOptions": {
        "compactGroups": false,
        "componentArrangementPolicy": "topmost",
        "edgeBends": true,
        "edgeBundling": true,
        "edgeToEdgeDistance": 5,
        "minimumLayerDistance": "auto",
        "nodeToEdgeDistance": 5,
        "orthogonalRouting": true,
        "preciseNodeHeightCalculation": true,
        "recursiveGroupLayering": true,
        "separateLayers": true,
        "webWorkers": true,
        "nodePlacer": {
            "barycenterMode": true,
            "breakLongSegments": true,
            "groupCompactionStrategy": "none",
            "nodeCompaction": false,
            "straightenEdges": true
        }
    }
}

For other out-of-the-box relations, go to Out-of-the-box relation types.

Tip Collibra offers several feature to help you create the semantic layer. For information, go to About the Semantic Layer submenu in Stewardship

About the conceptual layer

The conceptual layer is the highest level of organization in the Guided Stewardship operating model. It represents the overarching, context-independent data structures within an organization’s data landscape. It is where you can define concepts, such as Customer and Product, and their component fields, without direct reference to system-specific implementations.

The conceptual layer is closely related to the semantic layer. The main difference between them is that the conceptual layer is context-independent, whereas the semantic layer describes the structure in an individual System. The definition of the conceptual layer can be very useful for larger organizations that have data governance experience. For other organizations, defining the physical layer and the semantic layer can be a good starting point. If the full Guided Stewardship operating model is too complex for your organization, we advise to focus on the semantic layer because the semantic layer provides better out-of-the-box features and relation possibilities to the physical layer.

The organization of the conceptual layer is based on many-to-many relationships, which makes the conceptual layer more concise and flexible than tree-like arrangements that rely strictly on one-to-one and one-to-many relationships.

Show some examples

In this example, 3 Lines of Business assets group both the Customer and Product data domains. Each Data Domain groups several Data Concepts. In this case, both Data Domains group the Name and ID Number Data Concepts. This is possible because Name and ID Number, as Data Concept assets, represent the concepts abstractly, not as specific implementations. A Customer can have a Name and a Product can have a Name, but you need only one Data Concept asset to represent the idea of “name”.

In this next example, the Data Attribute Address is used in several Data Entities, such as Customer Address, Event Address and Employee Address.

Technology assets in the Guided Stewardship operating model

The following Technology assets are considered part of the Guided Stewardship operating system:

• System, which is part of the semantic layer.
• Database, which is part of the physical layer.

Related topics

• About Guided Stewardship
  
• About the Semantic Layer submenu in Stewardship
  
• Guided Stewardship diagram views
  
• For other possible out-of-the-box relation, go to Out-of-the-box relation types