integration layer

An integration layer is an architectural tier in an information system that connects, mediates, and coordinates data and processes between otherwise separate applications or platforms. In industrial and manufacturing environments, it commonly sits between shop-floor systems (for example MES, SCADA, PLC networks) and higher-level business or enterprise systems (for example ERP, quality, or planning tools).

The integration layer focuses on how systems communicate, not on executing production itself. It typically provides routing, data transformation, protocol adaptation, and orchestration of messages or events, so that each connected system can continue to use its own internal data structures and technologies.

Key characteristics

• Mediation and routing: Receives data or events from one system and delivers them to one or more target systems based on defined rules.
• Data transformation: Maps and converts data formats, codes, and structures (for example, aligning equipment IDs or material codes) so that systems with different schemas can interoperate.
• Protocol and interface adaptation: Bridges different technical interfaces, such as translating between OPC UA, message queues, REST or SOAP APIs, and file-based exchanges.
• Orchestration and workflows: Coordinates multi-step interactions across several systems, such as synchronizing production orders between ERP and MES and updating quality records.
• Centralized governance: Provides a controlled point for managing versions, monitoring integrations, and enforcing security and access rules across connected systems.

Operational context in manufacturing

In a manufacturing setting, the integration layer often appears as middleware platforms such as enterprise service buses, iPaaS tools, or custom API gateways. It may handle scenarios like:

• Exchanging production orders and confirmations between ERP and MES.
• Collecting equipment data from OT systems and publishing it to historians, analytics platforms, or operations intelligence tools.
• Synchronizing master data, such as materials, routes, and work centers, across multiple plants or systems.
• Standardizing metric definitions (for example, OEE, availability, or scrap) into consistent, site-specific schemas before they are consumed by reporting or KPI dashboards.

The integration layer does not replace MES, ERP, PLCs, or quality systems. Instead, it provides a structured way for them to exchange data reliably and consistently, especially when different plants or vendors use differing standards or versions.

Relation to standards and data contracts

When organizations align to standards such as ISA-95 or ISO 22400, the integration layer is often where those reference models are translated into concrete APIs, message schemas, and data contracts. The layer can:

• Implement canonical data models derived from standards while allowing each system to maintain its native model.
• Version and document how each metric or object is represented and computed across the environment.
• Provide a single place to adapt integrations when standards, system versions, or internal conventions change.

Common confusion

• Integration layer vs. MES or ERP: The integration layer does not plan, execute, or record production as a system of record. It only coordinates information flows between such systems.
• Integration layer vs. data lake or historian: A data lake or historian primarily stores and analyzes large volumes of data. The integration layer primarily moves and transforms data, though it may deliver data into these repositories.
• Integration layer vs. API: An API is a specific interface exposed by a system or the integration layer itself. The integration layer is the architectural tier that hosts, manages, and composes multiple APIs and other integration mechanisms.