Onibex Moves SAP ERP Data Into Real-Time Confluent Streams

Onibex is bringing SAP operational data into real-time streaming pipelines with OneConnect, its integration platform for SAP ECC, SAP S/4HANA, and other ABAP-based environments. The product connects those systems with Confluent Cloud, allowing historical records and ongoing business changes to move into Apache Kafka for use by analytics platforms, applications, and artificial intelligence workloads.

The approach addresses a recurring data-modernization problem. SAP systems contain the transactions behind sales, inventory, procurement, finance, and production, but downstream applications often receive that information through separate batch processes. By the time the data reaches a warehouse or dashboard, the underlying business event may already be hours old.

OneConnect Separates Data Access From ERP Migration

OneConnect uses application-layer change data capture to transmit SAP data as business records change. It can also load historical information in batches, allowing teams to establish an initial data set before activating real-time delta updates.

This architecture lets an SAP event serve more than one downstream consumer. A sales-order update, for example, can move into Kafka and then be used by analytics, customer applications, integration workflows, or machine-learning systems. Each consumer can use the same event stream without requiring a separate extraction process from SAP.

Support for both ECC and S/4HANA also separates the data-access decision from the ERP migration timetable. An organization does not have to complete its S/4HANA program before introducing streaming for a defined business requirement. It can begin with an existing ECC process and carry the integration pattern into the future landscape.

SAP Data Modeler Preserves Business Context

Moving data quickly is not enough if downstream teams cannot interpret it. OneConnect includes a graphical SAP Data Modeler for defining business entities from SAP tables, fields, and Core Data Services views. The modeler can incorporate standard structures and customer-created Z tables, giving organizations a way to represent their own SAP extensions.

Onibex also provides prepackaged SAP entity mappings. These reduce the need to reconstruct common objects such as sales orders, materials, customers, and vendors for every project. Teams can add fields, apply extraction filters, and control which data leaves SAP before it enters Kafka.

The platform converts SAP data types into the Avro format and can dynamically create the Confluent topics and schema-registry entries that define how messages are structured. Schema-evolution support is designed to keep downstream structures aligned as the source model changes.

Real-Time SAP Data Becomes Reusable Infrastructure

Once SAP data reaches Confluent, sink connectors can move it into destinations such as Databricks or Snowflake. Onibex presents this as a building-block model: companies can start with one sales, logistics, or analytics use case, then extend the same streaming foundation to additional consumers.

The larger shift is architectural. Instead of building a new extraction for each dashboard or application, SAP teams can publish governed business events once and reuse them across the enterprise. That makes real-time integration more than a reporting improvement. It becomes shared infrastructure for operational analytics, connected applications, and future AI workloads.

What This Means for SAPinsiders

• Streaming changes integration economics. When several applications consume the same SAP event stream, integration work becomes reusable infrastructure rather than repeated project spending. That can reduce duplication as new analytics and AI use cases emerge.
• Migration risk becomes easier to distribute. Establishing stable data streams before an S/4HANA cutover can separate downstream modernization from core ERP change. Teams can therefore avoid concentrating every integration dependency inside the migration program.
• Governance decisions move upstream. Modeling entities and applying filters before SAP data enters Kafka forces teams to define ownership and meaning earlier. This can reduce reconciliation work and conflicting definitions across downstream platforms.