Concept card

SAP Datasphere — Connections Catalogue

As of 2026-10-06

Edited by Cédric Mary, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

What is SAP Datasphere?

Whether a Datasphere use case is a 2-hour connector config or a 6-week custom integration project hinges entirely on which of the four connector categories the source system falls into.

What it is

The Datasphere Connections Catalogue is the set of source-system connectors Datasphere ships out of the box — the integration surface that lets a Datasphere space ingest, federate, or replicate data from anywhere the customer's data actually lives. For SAP analytics consultants, knowing what is in the catalogue (and what is not) decides whether a planned analytics use case is a 2-hour configuration or a 6-week custom integration project.

The four connector categories. SAP source connectors — S/4HANA Cloud + on-premise (via SAP Datasphere Agent), BW/4HANA, BW 7.x, ECC, HANA Cloud + on-premise, Ariba, Concur, SuccessFactors, Fieldglass; the SAP family is covered comprehensively, and SAP-to-SAP connections come with SAP-engineered semantic understanding (delta extraction, business content, hierarchy preservation). Cloud database connectors — Snowflake, Google BigQuery, Microsoft Azure SQL, Amazon Redshift, Databricks, MongoDB Atlas, PostgreSQL on cloud; one-click configuration with managed credentials. File-source connectors — Amazon S3, Azure Data Lake Storage Gen2, Google Cloud Storage, SFTP; for landing CSV/Parquet/JSON files into Datasphere. Application connectors — Salesforce, Workday, ServiceNow, Google Analytics 4, Adobe Analytics; primarily for marketing and HR analytics use cases not covered by SAP-source connectors.

Three usage patterns per connector. Remote table — the source is queried live, no replication; the federation pattern, suits low-frequency exploration. Replication flow — periodic full or delta replication of source data into Datasphere local tables; the operational pattern for high-frequency analytics. Data flow — ETL-style transformation pulling from source, transforming, landing in Datasphere local tables; for cases where the raw source data needs reshaping before analytics.

Why it matters

SAP-to-SAP connections carry SAP-engineered semantic understanding (delta extraction, hierarchy preservation) that generic cloud or file connectors don't.
Choosing remote table vs. replication flow vs. data flow trades live-query simplicity against performance and transformation needs.
Legacy on-premise databases (old Oracle, Informix, mainframe DB2) and niche SaaS applications aren't in the catalogue — they need Smart Data Integration adapters or a third-party ETL tool.

Key points

Four connector categories — SAP sources (S/4, BW, HANA, Ariba, SuccessFactors, etc.), cloud databases (Snowflake, BigQuery, Azure SQL, Databricks), file sources (S3, ADLS, GCS, SFTP), application sources (Salesforce, Workday, ServiceNow, GA4).
SAP-to-SAP connections — SAP-engineered semantic understanding (delta, business content, hierarchy preservation); the strongest connector class.
Three usage patterns — Remote Table (federation, live query), Replication Flow (periodic full/delta), Data Flow (ETL-style transform-on-ingest).
Not in catalogue — older on-premise databases (Oracle 8i, Informix, DB2 mainframe), specialised SaaS; need SDI custom adapters or external ETL.
First-question decisive — knowing the catalogue cold answers 'can we connect to X?' in 30 seconds, sets project schedule correctly.
The Connections Catalogue answers a data-plane question (can Datasphere reach this source); whether Joule or an agent can ACT on that source is a separate, agent-plane question governed by SAP AI Agent Hub.
Non-SAP-source connectors (application, cloud-database, file) carry no inherited Business Content — budget the same catalog/glossary work (C127) an AI grounding use case needs, in addition to the analytic modelling effort this card already flags.
Connector category determines modelling effort; usage pattern determines operational cost and latency; whether an AI use case is in scope determines whether catalog/access-control work must happen before go-live.

Terms used on this page

Connections catalogue: The set of source-system connectors Datasphere ships out of the box — SAP sources, cloud databases, file sources, application sources; the integration surface that determines what data Datasphere can natively reach.
Datasphere Agent: An on-premise component that lets Datasphere connect to on-premise SAP systems (S/4HANA on-prem, BW 7.x, HANA on-prem) and on-premise databases through a customer-controlled gateway; required for any non-cloud SAP source.
Smart Data Integration (SDI): HANA's custom-adapter framework — the fallback for sources not in the Datasphere connections catalogue; lets developers write custom adapters in C++ or use community adapters.
Business Content: SAP-pre-built data models, extractors, and transformations for SAP source systems; the SAP-to-SAP connections inherit Business Content semantics, dramatically reducing modelling effort for SAP-source analytics.
SAP AI Agent Hub: SAP's cross-vendor governance and inventory layer for agents, LLMs and MCP servers, discovering assets across Microsoft, Google, AWS, ServiceNow and SAP AI Core — a separate question from whether the Connections Catalogue can reach a given source system for data.
Data-plane vs. agent-plane question: The distinction between "can Datasphere ingest or federate this source's data" (Connections Catalogue, data plane) and "can an agent call this source's APIs to take an action" (SAP AI Agent Hub, agent plane) — two different integration surfaces a client often conflates in one question.
Generic OData connector: A Datasphere connector type for consuming any standards-compliant OData service that is not covered by a named SAP or application connector — the fallback for a source with an OData interface but no dedicated catalogue entry.

Sources

Full card available to members. What the full card adds: the full decision framework · the SAP vs Snowflake / Databricks / Fabric comparison · the common pitfalls and their fix · the cheat sheet · the architecture schemas · the code blocks · the facts worth quoting.

Open in the app →