Concept card

BDC Connect for Microsoft Fabric

As of 2026-07-23

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

What is BDC Connect for Microsoft Fabric?

BDC Connect for Fabric uses OneLake Mirroring to give near-real-time, read-only SAP data inside Power BI/Fabric without duplicating the corpus — solving the classic standoff where business units want live SAP data but IT won't open raw S/4HANA access.

What it is

BDC Connect for Microsoft Fabric is the SAP-sanctioned integration layer that lets SAP Business Data Cloud and Microsoft Fabric share governed data in both directions — without copy-and-forget pipelines. Announced in November 2025 (early access; GA originally targeted H1 2026 per the SAP roadmap published at Microsoft Ignite, then pushed to the second half of 2026 — SAP confirmed a Q3 2026 GA target at Sapphire in May 2026, still not shipped as of 2026-08-16), it solves the most common cross-vendor impasse in large-enterprise SAP shops: business units that have standardised on Power BI and Fabric want live SAP data, but the IT team refuses to hand over raw S/4HANA access.

The mechanism is OneLake Mirroring for SAP, with Datasphere acting as the source-side staging and governance layer. A Datasphere table or view surfaces as a queryable OneLake shortcut inside a Fabric Lakehouse — read-only, governed by SAP Data Access Controls, and refreshed automatically as the Datasphere source changes. In the reverse direction, a Fabric Lakehouse table registered as a Fabric Data Product appears inside BDC as a foreign Data Product, queryable from Datasphere views and accessible to Joule agents. The bidirectionality closes the loop: SAP governs the golden copy of S/4 actuals; Fabric hosts the Power BI reports and Copilot workloads; neither side needs to duplicate the corpus.

Why it matters

Latency is the explicit scoping variable: for an SAC Live use case already tolerating 15-minute lag, migrating to Power BI via BDC Connect can match freshness at a fraction of the CU consumption.
For guaranteed 5-second-latency Joule answers on transactional tables, the direct Datasphere-to-SAC path stays the right choice — mirroring adds hop overhead.
Pricing has no cross-vendor bundling — BDC Capacity Units plus Fabric F-SKU separately — and GA itself has slipped from the original H1 2026 target to the second half of 2026 (Q3 2026 per SAP's Sapphire confirmation), not yet live as of 2026-08-16 — both a cost and a timeline conversation to have explicitly with the client's finance team.

Key points

OneLake Mirroring for SAP: Datasphere tables appear as read-only OneLake shortcuts in Fabric — governed by SAP Data Access Controls, refreshed automatically.
Bidirectional: Fabric Data Products appear as foreign Data Products queryable from BDC, including by Joule agents.
Pricing is additive, not bundled: BDC Capacity Units (SAP side) + Fabric F-SKU (Microsoft side).
Latency: OneLake Mirroring is near-real-time (seconds to minutes) — not batch, but not sub-second Joule-grade either.
Best fit: enterprises mandating Fabric as the primary BI layer who want governed SAP actuals without migrating the Power BI estate to SAC.
Avoid when: real-time Joule grounding on S/4 transactional tables is the primary need, or the Fabric estate does not yet exist.
The governance boundary is clear: SAP Data Access Controls on Datasphere govern what rows the Fabric consumer sees — one rule set, two surfaces.
Early access announced November 2025 at Microsoft Ignite; GA originally targeted H1 2026, since pushed to the second half of 2026 (Q3 2026 confirmed at SAP Sapphire, May 2026) — not yet GA as of 2026-08-16.

Terms used on this page

OneLake Mirroring for SAP: Microsoft Fabric feature that replicates SAP Datasphere tables into Fabric OneLake as read-only shortcuts, refreshed continuously — the transport layer underlying BDC Connect for Fabric.
Fabric F-SKU: Microsoft Fabric capacity unit (F2 through F2048) — the compute unit that prices all Fabric workloads including OneLake storage, Power BI, Copilot, and Data Engineering. F64 is the practical minimum for enterprise-grade SAP mirroring.
OneLake shortcut: A Fabric feature that makes an external data source (Azure Data Lake, S3, SAP Datasphere, etc.) appear as a virtual folder inside a Fabric Lakehouse, without copying the data into Fabric storage.
Foreign Data Product: In BDC terminology, a data product originated outside SAP (e.g. a Fabric Lakehouse table) that is registered in BDC's data catalog and made queryable from Datasphere views and Joule agents.
Data Access Control: Datasphere's catalog-level row and column security mechanism — a single rule that governs access across SAC, Joule, Delta Sharing, OData, and mirrored Fabric consumers simultaneously.

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.

Open in the app →