For non-RISE SAP customers (on-premise S/4HANA and ECC) building AI, analytics, and automation on cloud targets like Databricks, Snowflake, Fabric, watsonx, or SAP BDC.
Three quiet changes that made cloud data integration harder, not easier.
Live, semantic, secured, standard. Four properties, one integration surface.
450K+ business-named SAP objects (Purchase Order, Sales Invoice, Delivery, and so on) exposed as the primary interface. Raw table and view access with delta CDC is available as an option, not the default.
450K semantic + 850K tables & views = 1.3M+ SAP objectsPulled live via API into any downstream app or open storage (data lake, NFS). No batch window, no intermediate warehouse.
0.143s on a 100M+ record S/4HANA tableNative SAP authorizations are carried on every access. No parallel credential model to build, audit, or reconcile against your identity provider.
Clean-core SAP objects exposed through a standard REST API. Does not rely on the ODP interface. SAP-certified and Microsoft-certified.
What SAP's BDC integration entitlement looks like today, and where SIP fits.
| RISE customers | Non-RISE customers | |
|---|---|---|
| Who is impacted | S/4HANA Cloud Public Edition (GROW), Private Edition (RISE), and on-premise systems migrated under a RISE contract. | S/4HANA on-premise without a RISE contract, and SAP ECC. |
| Semantic models / data products | SAP-managed, curated, governed data products, ready to use. | Not provided. Customer must build their own data products. |
| Ecosystem access | Databricks, Snowflake, Google Cloud, Collibra, Confluent, DataRobot, Microsoft Fabric. | Can onboard to BDC (Datasphere object store + data integration capacity), but without SAP-managed content. |
| Data sharing model | Zero-copy, bi-directional, business context preserved. | Customer-built pipelines. No SAP-managed zero-copy entitlement. |
| What is needed to close the gap | Nothing. Included in the subscription. | Semantic modeling, extraction, and governance. Done in-house, by a consultant, or by SIP. |
Streaming access for AI, analytics, and automation is the top data-integration priority in most SAP shops. The market for non-SAP sources is well served. For SAP, three specific constraints have narrowed the options for non-RISE customers:
SIP clears all three. Legally, in real time, and with business semantics preserved. It hands off clean, business-named data (or raw tables, with CDC) to whatever comes next downstream:
Patented SIP technology, available across the environments you already use.
Azure, AWS, GCP, or on-premise. Same integration surface across all four.
SAP-certified. Compatible with SAP BAIP (formerly BTP), available since 2022.
Standard REST API. Usable in SAP Datasphere, other SAP apps, or any non-SAP AI, analytics, or automation platform.
The questions we hear on most first calls with non-RISE data and platform teams.
RISE is a contract decision that touches licensing, infrastructure, operations, and roadmap timing. Data access is one input into that decision, not the whole decision. Many customers have valid reasons to stay non-RISE for the near term: on-premise control, sovereignty, existing infrastructure investments, cost profile, or timing on S/4HANA readiness. SIP lets the data-integration decision proceed on its own timeline, whichever way the contract decision eventually goes.
SAP Note 3255746 restricts third-party use of the ODP API for real-time extraction from on-premise and Private Cloud Edition sources. SIP does not use the ODP interface. It exposes SAP business objects through a standard REST API on clean-core SAP objects, which is a different integration surface. That is the structural reason SIP is unaffected by the note.
Yes. The default interface is semantic (business-named objects), but table-level access with delta CDC is available as an option. Most customers use semantic for AI and analytics workloads and reserve table-level CDC for specific replication cases.
More than 1.3 million. That breaks down into 450K+ business-named semantic objects (the default interface) and 850K+ underlying SAP tables and views (available as an option, with delta CDC). Same platform, two access modes.
Any REST-consuming target. In production today: SAP Datasphere, SAP BDC, Databricks, Snowflake, Microsoft Fabric, watsonx, Confluent, and open storage (data lake, NFS). Custom AI or analytics apps consume the same API.
Native SAP authorizations are carried on every access. There is no parallel credential model to build, audit, or reconcile against your identity provider. That preserves the role-based access model your security team already governs.
0.143 seconds, measured on a table with more than 100 million records in S/4HANA. That is the platform performance figure; end-to-end latency depends on your downstream target.
More questions on architecture, certification, security, and deployment:
Read the full FAQ →The STREAM Engine is the zero-copy, SAP-certified streaming core that powers every SIP integration, non-RISE and RISE alike.