Skip to main content

14.5 — Sharing, Exposure & Consumption Security

1. Security does not stop at the space​

A secure model can become insecure through an incorrectly designed sharing or consumption boundary.

Use this chain:

SOURCE | SPACE | MODEL | SECURITY | SHARE / EXPOSE | CONSUMER

Every boundary requires an explicit decision.

2. Sharing versus exposure​

Sharing makes data available across Datasphere space boundaries.

Exposure makes selected data available for consumption outside the modeling context.

They solve different problems.

3. Share the smallest useful object​

Prefer:

  • governed reusable models;
  • curated dimensions;
  • semantic products;
  • approved analytical models.

Avoid exposing raw acquisition objects unless there is a strong reason.

4. Consumption security​

Consumers may include:

  • SAP Analytics Cloud;
  • Excel-based analytical clients;
  • external applications;
  • SQL clients;
  • other approved consumers.

The consumer should receive only the data surface required for its purpose.

5. Analytical security​

Security must remain correct when the consumer aggregates data.

Example:

If a user can see only Germany, an analytical query must not reveal France indirectly through totals, drill paths, or unrestricted dimensions.

This requires testing the complete consumption path.

6. Cross-space architecture​

Example:

PROCUREMENT RAW | v PROCUREMENT CURATED | v ENTERPRISE SEMANTIC | +------> SAC | +------> ANALYSTS

Each arrow should have a documented security contract.

7. External exposure review​

Before exposing data ask:

  1. Who consumes it?
  2. Which authentication path is used?
  3. What object is exposed?
  4. Does row-level security still apply?
  5. Can metadata reveal sensitive information?
  6. Is write access possible?
  7. What audit evidence exists?
  8. How is access revoked?

8. Security contract​

For every published data product define:

  • owner;
  • classification;
  • consumer;
  • allowed purpose;
  • security model;
  • refresh;
  • retention;
  • audit;
  • revocation;
  • change policy.

Architect exercise​

Design a Finance analytical product consumed by:

  • CFO;
  • regional finance leaders;
  • country controllers;
  • auditors.

Defend the difference between model access, data access and consumption access.