Skip to content

CONCEPT Cited by 1 source

Semantic layer of business concepts

Definition

A semantic layer of business concepts is an intermediate layer — between source-system schemas and a target schema (such as the logical data model of a golden record) — that captures the business vocabulary of the domain: what concepts exist, their attributes, and their relationships.

In Zalando's MDM framing:

"By using knowledge graphs for a live-data representation of all systems' logical data models and how they map to a semantic layer of business concepts, we are able to automatically generate the logical data model of the golden record…" — Zalando MDM (sources/2021-07-28-zalando-knowledge-graph-technologies-accelerate-and-improve-the-data-model-definition)

What sits in the semantic layer

Three node types in Zalando's knowledge graph:

  • Concept — a business concept (e.g. Address, Business Partner). Keyed entity classes.
  • Attribute — a typed property of a concept (e.g. street name on Address).
  • Relationship — a directed, typed edge between two concepts (e.g. Business Partner "has contact" Address).

Every source-system column is mapped into this layer, not directly into the target schema. The target schema (logical data model) is then derived from the semantic layer.

Why a semantic layer

Without it:

  • Source → target mappings are direct and hard-coded; refactoring the target schema forces every source mapping to be re-derived.
  • Domain experts have no common vocabulary to discuss with engineers — the "exact same name for concepts, attributes, and relationships" problem Zalando names explicitly.
  • Multiple source-system columns representing the same business fact have no shared home.

With it:

  • Source mappings target a stable business vocabulary; the target schema can evolve independently as long as the derivation rules stay consistent.
  • Business-engineering dialogue has a shared artifact: "the domain expert can provide us with these definitions and some coordination that the exact same name for concepts, attributes, and relationships is required. This is done by cross-referencing system's business concepts and unifying their wording."
  • Multiple source columns representing the same business fact ( System A's three address_line_N columns vs. System B's structured fields) all converge on the same Concept / Attribute nodes.

Relation to domain modeling

Zalando's semantic layer is a narrower, MDM-focused cousin of the broader domain model concept used in enterprise data-integration platforms like Netflix UDA. Both treat business concepts as first-class entities stored as data. Zalando's implementation is a Neo4j property graph at design-time scale; Netflix UDA uses RDF + SHACL + named graphs at enterprise scale. Both serve semantic interoperability — the property that downstream consumers agree on what data means, not just how it is encoded.

Seen in

Last updated · 476 distilled / 1,218 read