Concept card

Delta Sharing Protocol

As of 2026-07-24T14:00:00Z

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 Delta Sharing Protocol?

Delta Sharing's zero-copy, client-agnostic design lets BDC participate in the open data ecosystem on equal terms with Snowflake and Databricks — the protocol, not a proprietary connector, is what makes cross-tool interop work.

What it is and why it matters

Delta Sharing is an open protocol for sharing tables in the Delta Lake format across organisations, tool stacks, and cloud regions — without copying the data and without locking either side into a single vendor. It began life as a Databricks project and now lives under Linux Foundation governance, which is precisely why it has been adopted well beyond Databricks: Snowflake, Iceberg-based lakehouses, and SAP itself all speak the same protocol. For SAP Datasphere and SAP Business Data Cloud, Delta Sharing is the bridge that lets a customer's data leave, or enter, the SAP estate without a bespoke integration project. A share is defined once, a recipient is issued scoped, revocable credentials, and from that point on the data moves — or rather, does not move — on demand.

The reason this matters commercially as much as technically: most cross-organisation and cross-platform data exchange used to require either a point-to-point extraction pipeline, a VPN tunnel into a shared database, or a manually refreshed file drop. All three carry ongoing engineering cost, security surface, and staleness risk. Delta Sharing collapses that into a protocol-level concern — the producer publishes once, and any number of consumers subscribe using whatever tool they already run.

How it works

Why it matters

Zero-copy sharing means data moves only at query time via pre-signed URLs — no ETL pipeline or shared storage negotiation needed for B2B data exchange.
Client-agnostic consumption (Power BI, Tableau, Spark, pandas can all read a share) means a Datasphere-published share reaches teams on any tool without schema drift.
Credential-scoped, revocable access lets a producer share with one named recipient and cut access instantly — a governance lever a generic file export can't match.

Key points

Open Linux-Foundation protocol; Datasphere is producer + consumer; cross-cloud + cross-tool.
Three layers: share definition · recipient credentials · pre-signed file URLs (zero-copy).
DAC propagates from Catalog onto shares automatically — moat vs Snowflake/Databricks where re-implemented.
Producer-side: zero compute impact; consumers read object store directly.
Four use cases: B2B exchange · marketplace consumption · cross-tool interop · multi-region replication-light.
Replicate-vs-share threshold: ~100 GB/partner-month; below shares win on TCO, above replicate.
Token rotation quarterly default, monthly for regulated industries; SIEM wiring mandatory for audit.
Schema-narrowing or column removal = contract change; backwards-compatible changes non-breaking.
Delta Sharing Protocol 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

Delta Sharing: Open protocol for sharing Delta tables across orgs and tools. Linux Foundation OSS, Databricks-authored.
Share definition: Named bundle of tables a producer makes available, with column-projection and row-filter options.
Recipient credential: Token-protected URL the producer issues to a specific consumer organisation. Bearer-token semantics.
Pre-signed URL: Time-limited URL pointing directly to Delta files in producer object store. Consumer reads bytes without producer compute.
Zero-copy share: Consumer reads producer-stored files at query time; no replication or pre-staging required.
Recipient lifecycle: Provision → rotate → audit → revoke flow per partner relationship.
Time-travel query: Point-in-time read of a shared table at a specific version, for audit reproducibility.
Iceberg sharing: Apache OSS alternative protocol; semantically similar to Delta Sharing. Preview in Datasphere 2026.

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 →

Guides that answer with this page

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