Concept card

Live vs Import Connections

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 Live vs Import Connections?

The 2026 senior default is live-to-Datasphere-AM for everything — imported models drop DAC inheritance entirely, meaning every story viewer sees every row regardless of permission.

Should I use a live or import connection in SAP analytics cloud?

Short answer above

What it is

Every SAC story binds to data via either a live connection (query-on-render against the source) or an imported model (snapshot copied into SAC at refresh time). The choice determines governance, freshness, performance, and cost — and gets re-litigated on every project. The senior default in 2026: live to a Datasphere analytic model. Everything else needs a documented reason in the ADR.

Live connection economics. Each render queries the underlying analytic model (or BW InfoCube, or HANA Calc View). DAC inheritance flows through automatically; users see only what their permissions allow. Lineage is automatic; the catalog tracks the dependency. Latency = AM query time + network. For governed enterprise reporting, this is the canonical path.

Imported model economics. Data is copied into SAC at scheduled refresh. Render is fast (data is local). DAC does NOT inherit — every user with story access sees every imported row. Lineage breaks at the import boundary. The model goes stale until next refresh. Tolerable only for one-off exploratory analysis on non-governed data.

The 4-axis decision matrix. Governance: live wins (DAC + lineage). Freshness: live wins (always current). Performance: import wins for high-frequency dashboards on small data; live wins everywhere else. Cost: live runs AM queries (CU spend); import copies storage (storage spend) — usually a wash, but live offers better governance per euro.

Why it matters

Import breaks DAC inheritance completely — anyone with story access sees every imported row, a hard fail against GDPR Art. 5 and EU AI Act Art. 10 in regulated domains.
Mixing live-BW and live-Datasphere connections in the same story creates governance drift where users can't tell which dimension is actually governed.
Live wins governance and freshness by default; import only wins on raw render performance for high-frequency dashboards on small, non-governed data.

Key points

Live = governed (DAC + lineage); import = fast but ungoverned
Default 2026: live to Datasphere analytic model
BW migration 1-3 stories/quarter from 2026
Imported tolerable only for one-off exploratory work
Mixed live-BW + live-DSP in one story = drift
DAC re-verify on every rebind
Quarterly connection audit retires unused
Imported flat in regulated domains = GDPR + AI Act fail
Live vs Import Connections is mastered only when it changes a named buyer decision.
Start with the semantic contract and control model before demonstrating the tool.

Terms used on this page

Live connection: Story queries source on render; data never copied into SAC
Imported model: Source data snapshot copied into SAC at refresh time
DAC inheritance: Row+column permissions flow from source to consumption
Federation: Live query against on-prem or external source via Cloud Connector
Refresh schedule: Cadence at which imported model re-snapshots from source
Cloud Connector: SAP-managed gateway for cloud-to-on-prem connectivity
Lineage: Tracked dependency from source to consumption surface
Catalog connection audit: Quarterly retire-unused review of stored connections

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 →

Guides that answer with this page

These guides cite this page as one of the sources their answer rests on.