BDC Connect: A direct link to Databricks, Snowflake, and others

Table of contents 

1. The Problem: The Dead End of Legacy Architectures

Historically developed integration approaches suffer from significant efficiency losses. Traditional ETL processes require data to be physically copied, which not only consumes storage space but also jeopardizes the integrity of the information.

A key problem is “semantic collapse”: When SAP tables are extracted in isolation, valuable metadata—such as hierarchies, currency conversion logic, and business rules—is lost. This information must be painstakingly reconstructed in the target platforms, which often leads to inconsistencies. In addition, delayed replication means that decisions are made based on outdated data.

2. Strategic Connectivity: Partners and Roadmap

SAP BDC Connect is designed as an open architecture that connects various ecosystems. This enables companies to use their preferred analytics platform while SAP retains data ownership. The platforms are integrated either as OEM solutions (deeply integrated) or as Solution Extensions (SOLEX).

3. Technical Deep Dive: The Zero-Copy Delta Sharing Protocol

At the heart of BDC Connect is the Delta Sharing Protocol. It is the first open standard for secure data sharing that ensures a high degree of interoperability across different cloud platforms. Unlike traditional interfaces, this protocol operates directly at the storage level of the cloud object storage.

The process begins with a metadata handshake during which schema and partition information is exchanged. Instead of physically moving the data, the BDC Connect Server generates short-lived, cryptographically secured links (pre-signed URLs) that point directly to the underlying storage. The target system’s compute engine uses these URLs to read the data packets in parallel and at maximum bandwidth. This eliminates the need for an intermediate integration server, which increases scalability almost indefinitely.

This is complemented by the ORD (Open Resource Discovery) and CSN (Core Schema Notation) standards. These ensure that the target platform not only receives raw data but also understands its meaning. Fields are automatically typed, relationships are recognized, and business descriptions are adopted.

4. Terms of Use: Rights and Restrictions

To ensure compliance and architectural performance, SAP has defined clear guidelines for the use of BDC Connect. These guidelines govern the handling of the data products provided in third-party systems.

It is important to note that transformed or enriched data is considered a new, independent entity. Once the original SAP data has been modified by linking it to external data or through AI-driven processing, it is no longer subject to materialization restrictions and can be permanently stored in the target system.

5. Identity Management and Universal Governance

End-to-end security is a key priority for CIOs and CSOs. BDC Connect addresses the issue of fragmented security approaches by providing an end-to-end identity chain. In this setup, various directory services and authentication protocols work together seamlessly.

Microsoft Entra ID serves as the primary source for all user identities. This information is automatically synchronized to the SAP Cloud Identity Service (IAS) via the Identity Provisioning System (IPS). The key advantage of this architecture lies in its integration with catalogs such as the Unity Catalog or Snowflake Horizon. Since these systems use the same identities, granular access rights can be defined at the row and column level, which apply consistently across the entire hybrid landscape.

6. Cost Model and Cost-Effectiveness

The pricing structure for SAP Business Data Cloud Connect is modular and based on a three-pillar model:

1. Capacity Units for base licensing,
2. Infrastructure fees for technical operations and
3. Premium fees for specialized partner workloads. 

To assess the cost-effectiveness of a project, it is essential to understand which system triggers which fees and when.

6.1. Commercial Comparison of Partner Models

Depending on how a third-party system is integrated (as a core component, solution extension, or external connector), different cost structures apply. The following table indicates whether infrastructure fees or premium fees apply.

6.2. Detailed Fee Structure

Infrastructure fees are charged for the technical operation of the data exchange. These fees consist of costs for data transfer and costs for processing requests.

• Inter-Region Data Transfer: This fee applies only when data is transferred from SAP BDC across hyperscaler boundaries (e.g., from Azure to AWS) or regional boundaries (e.g., from EU-West to US-East). Data transfers within the same region on the same hyperscaler are generally free of charge.
• Processing Requests: These are API calls. Every request made by the target system to the BDC Connect Server to retrieve data is counted.

6.3. Premium Fees and Compute Uplift

The Premium Fee is charged to selected partners (currently Snowflake and Microsoft Fabric). It is a surcharge based on the computational load that the SAP data generates in the target system.

For example, the process for calculating the premium fee at Snowflake works as follows: Snowflake monitors the usage of SAP data products and identifies the computational load (workload) directly associated with this data. Periodically, Snowflake reports the net price the customer paid for these specific workloads to SAP. SAP uses this information as a basis to calculate an uplift in BDC Capacity Units, which is deducted from the customer’s quota. Depending on the BDC discount tier, this effective uplift is expected to exceed 40 percent of the computing costs on the partner side.

7. Case Study: Developing a Cash Flow Forecast

The benefits of this architecture are clearly illustrated by the example of a liquidity forecast.

This scenario illustrates the process from raw data provision to analytical value:

1. Data Provision: S/4HANA provides relevant financial data as a semantically rich data product.
2. Zero-copy access: Through BDC Connect, the Databricks environment gains immediate live access to this data without any replication.
3. Advanced Analytics: A data scientist trains an ML model in Databricks that enriches SAP data with external market data.
4. Feedback loop: The forecasting results are sent back to the Business Data Cloud as an enriched data product.
5. Basis for decision-making: Management seamlessly accesses the final insights in SAP Analytics Cloud (SAC) based on a real-time data source.

8. Conclusion and Outlook

SAP BDC Connect offers the ideal balance between flexibility, control, and cost-efficiency. It enables companies to combine their investments in SAP software with the capabilities of modern cloud innovations. While specialized OEM solutions such as SAP Databricks offer the deepest level of integration, BDC Connect ensures alignment with a universal enterprise strategy. The future focus is on supporting additional data formats such as Apache Iceberg to further enhance the platform’s openness.