Terminology of derived relations

Term	Definition
Knowledge graph	Knowledge graph is a semantic metadata layer that maps relationships between technical data assets, their business context, applicable policies, and the people who use them. It provides a unified, searchable map of an organization's entire information landscape.
Relation type	A relation type indicates a relation's meaning. It specifies between which asset types a relation can be established. It has a contract defined by the head, tail, role, and co-role. Although explicit relation types and derived relation types are defined differently, both are based on the same type of contract.
Head (or source) and tail (or target)	The head and tail identify the types of assets that can be linked through a relation type. In the REST API, the head is called the source and the tail is called the target. This alternative terminology may sometimes be visible in the user interface.
Role and co-role	A role indicates a relation’s meaning when navigating from the head to the tail, whereas a co-role provides that meaning when navigating from the tail to the head.
Relation direction	A relation can be navigated either from the head to the tail through the role, or from the tail to the head through the co-role.
Asset type assignment	For a relation to be created or derived between two assets, the relation type needs to be first assigned to the head or tail asset types. The relation type can also be assigned to the descendants (subtypes) of the head or tail asset types.
Asset layout	After a relation type is assigned to an asset type, the relation type is automatically shown on the associated asset pages. If, however, the asset type has an out-of-the-box or custom layout, the relation type needs to be manually added to the asset layout.
Explicit relation type (ERT)	An explicit relation type defines the meaning of a direct link between two asset types. An explicit relation type is what you have always known as a relation type in Collibra.
Derived relation type (DRT)	A derived relation type defines a navigable path through the knowledge graph by following explicit relation types from one asset type to another asset type. This navigation is defined through one or more relation paths. A relation path may include other derived relation types.
Relation path	A relation path is a list of linked relation types that begin with an asset type and lead to another asset type. In a relation path, each relation type is added with a specific direction to ensure that the end of a relation type is compatible with the beginning of the next relation type. A relation path can contain both explicit and derived relation types. The asset type at the end of a relation type must always be compatible with the asset type at the beginning of the next relation type.
Relation	A relation is an instance of a relation type. It is an actual connection between two assets. It represents how the two assets are linked: either as an explicit relation or as a derived relation.
Explicit relation	An explicit relation is an instance of an explicit relation type. It is an actual connection between two assets that are directly linked. An explicit relation is implemented by physically registering the link in the database. Because it is a physical link, you can manually create an explicit relation. An explicit relation is what you have always known as a relation in Collibra.
Derived relation	A derived relation is an instance of a derived relation type. It is an actual connection between two assets that are indirectly linked. A derived relation is dynamically calculated and isn't stored in the database. Because it is a calculated link, you can't manually create a derived relation. It is a read-only result on an asset page that helps you understand how the asset is linked to other assets through a path of explicit relations.
Node	In a knowledge graph, assets are considered nodes.
Edge	In a knowledge graph, explicit relations are considered edges. Derived relations aren't strictly considered edges because they are dynamically calculated and aren't stored in the database.

Knowledge graph

Knowledge graph is a semantic metadata layer that maps relationships between technical data assets, their business context, applicable policies, and the people who use them. It provides a unified, searchable map of an organization's entire information landscape.

Relation type

A relation type indicates a relation's meaning. It specifies between which asset types a relation can be established. It has a contract defined by the head, tail, role, and co-role.

Although explicit relation types and derived relation types are defined differently, both are based on the same type of contract.

Head (or source) and tail (or target)

The head and tail identify the types of assets that can be linked through a relation type.

In the REST API, the head is called the source and the tail is called the target. This alternative terminology may sometimes be visible in the user interface.

Role and co-role

A role indicates a relation’s meaning when navigating from the head to the tail, whereas a co-role provides that meaning when navigating from the tail to the head.

Relation direction

A relation can be navigated either from the head to the tail through the role, or from the tail to the head through the co-role.

Asset type assignment

For a relation to be created or derived between two assets, the relation type needs to be first assigned to the head or tail asset types. The relation type can also be assigned to the descendants (subtypes) of the head or tail asset types.

Asset layout

After a relation type is assigned to an asset type, the relation type is automatically shown on the associated asset pages. If, however, the asset type has an out-of-the-box or custom layout, the relation type needs to be manually added to the asset layout.

Explicit relation type (ERT)

An explicit relation type defines the meaning of a direct link between two asset types.

An explicit relation type is what you have always known as a relation type in Collibra.

Derived relation type (DRT)

A derived relation type defines a navigable path through the knowledge graph by following explicit relation types from one asset type to another asset type. This navigation is defined through one or more relation paths. A relation path may include other derived relation types.

Relation path

A relation path is a list of linked relation types that begin with an asset type and lead to another asset type. In a relation path, each relation type is added with a specific direction to ensure that the end of a relation type is compatible with the beginning of the next relation type.

A relation path can contain both explicit and derived relation types. The asset type at the end of a relation type must always be compatible with the asset type at the beginning of the next relation type.

Image of relation paths in a DRT

Relation

A relation is an instance of a relation type. It is an actual connection between two assets. It represents how the two assets are linked: either as an explicit relation or as a derived relation.

Explicit relation

An explicit relation is an instance of an explicit relation type. It is an actual connection between two assets that are directly linked.

An explicit relation is implemented by physically registering the link in the database. Because it is a physical link, you can manually create an explicit relation.

An explicit relation is what you have always known as a relation in Collibra.

Derived relation

A derived relation is an instance of a derived relation type. It is an actual connection between two assets that are indirectly linked.

A derived relation is dynamically calculated and isn't stored in the database. Because it is a calculated link, you can't manually create a derived relation. It is a read-only result on an asset page that helps you understand how the asset is linked to other assets through a path of explicit relations.

Node

In a knowledge graph, assets are considered nodes.

Edge

In a knowledge graph, explicit relations are considered edges.

Derived relations aren't strictly considered edges because they are dynamically calculated and aren't stored in the database.