Now accepting guest editorial from SAP, Oracle, Workday and ServiceNow practitioners.Pitch a guide
The Sign-Off Table

Lesson 4 of 6

Data provisioning at scale

10 min

Listen to this lesson

0:00 / 6:19

At enterprise scale, test data stops being a script problem and becomes a supply chain. Thousands of automated runs a week across S/4HANA, BTP and partner systems need valid, consistent, private-safe data, and they need it in the right environment at the right time. The course on automation data strategy (sap-tau-302) covered techniques inside a test: self-contained creation, pools and runtime discovery. This lesson designs the provisioning architecture around them.

Sources of test data

SourceStrengthsWeaknessesBest for
Synthetic, created through APIsNo personal data; repeatable; preciseEffort to model; may miss real-world odditiesAutomated regression; negative cases
Masked production copyRealistic volumes and edge casesPrivacy controls needed; large; refresh disrupts testsIntegrated testing; UAT; performance
Subset copyRealistic but smallerTooling and consistency rules neededProject systems with limited capacity
Migrated data (mock loads)Exactly what will go liveOnly available during migration wavesMigration and reconciliation testing
Sources of test data and their trade-offs.
The Sign-Off Table · thesignofftable.com/academy/sap-tau-401

Most enterprises combine them. Contoso's pre-production system CMR is refreshed from production every quarter and masked as a post-copy step; the automation suite creates its own synthetic customers and orders on top; and during Wave 2, mock loads of Alternadores del Norte data into company code 1910 give the migration team, and the automation team, real legacy data to test with: 1,946 in-scope customers out of 3,180, including the largest, Flotillas Monterrey S.A. de C.V., and roughly 2,600 core return agreements in the custom table carried over from ECC.

A provisioning architecture

CatalogA register of data sets and generators: what exists, for which processes, in which environments, with which privacy status.
GeneratorsData factories that create synthetic master and transactional data through released APIs and BAPIs.
Copies and maskingRefresh and client copy procedures run by Basis, with masking rules agreed with process owners and data protection.
Pools and reservationsReady-to-use consumable data with atomic reservation for parallel runs.
Provisioning serviceOne interface tests call: "give me a fleet customer in 1710 / 10 / 00 with credit limit above $50,000".
Layers of a test data provisioning architecture.
The Sign-Off Table · thesignofftable.com/academy/sap-tau-401

The provisioning service is the key design idea. Tests should not know whether their data was created synthetically, taken from a pool or found in a masked copy. They ask for data by characteristics; the service decides how to satisfy the request in the current environment. When the environment changes, for example after a refresh, only the service's strategy changes, not hundreds of tests.

Tests ask by characteristics; the provisioning service chooses pool, generator or copy for the current environment.
The Sign-Off Table · thesignofftable.com/academy/sap-tau-401

Copies, clients and refreshes

SAP offers several copy mechanisms with very different scopes. A full system copy replaces everything, including users and configuration. A client copy copies one client's data within or between systems. Subset and scrambling tools copy selected data with transformations; SAP has offered such capabilities in its test data tooling over the years, so check what is licensed and supported in your landscape and under your RISE contract. Whatever the mechanism, three rules hold.

Copy mechanisms differ in scope, and every one of them is followed by masking and an automation re-seed.
The Sign-Off Table · thesignofftable.com/academy/sap-tau-401
  • Schedule copies in the shared environment calendar and re-seed automation data immediately afterwards.
  • Mask before anyone, including automation, uses the copy; verify masking with key processes.
  • Record which copy and which masking version an environment holds, so results can be interpreted later.

Scale and consistency across systems

Cross-system tests need consistent data in every system they touch. A fleet customer created in S/4HANA must exist with the same identifier in the BTP warranty portal and in the virtual EDI partner's configuration. Design generators to create cross-system sets in one call and to record the identifiers in every system, and design replenishment so that pools in different systems stay in step.

  1. RequestThe test asks for a fleet customer with an active warranty contract.
  2. S/4HANABusiness partner created through the API with sales area 1710 / 10 / 00 data.
  3. BTPWarranty contract created in the extension through its own API, referencing the business partner.
  4. Virtual partnerEDI partner configuration registered for the new customer number.
  5. ReturnAll identifiers returned to the test and written to the run log.
Provisioning one consistent data set across three systems.
The Sign-Off Table · thesignofftable.com/academy/sap-tau-401

What would you do?

Contoso's nightly suite grows to 900 tests across four teams. Each team has its own data scripts, and teams keep consuming each other's records. What should the architect introduce?

Key points
  • Combine synthetic, masked copy, subset and migrated data; each has a role.
  • Build provisioning as layers: catalog, generators, copies and masking, pools, and a provisioning service.
  • Tests request data by characteristics; the service decides how to satisfy them per environment.
  • Record which copy and masking version each environment holds; mask before use.
  • Create cross-system data sets consistently and plan volume data with Basis.

Check your understanding · not graded

Lesson check

  1. 1. What is the main benefit of a provisioning service that tests call by data characteristics?

  2. 2. Which data source best supports migration and reconciliation testing for Wave 2?

  3. 3. What should be recorded for each environment after a copy?

  4. 4. A cross-system test needs a fleet customer in S/4HANA, the BTP warranty portal and the virtual EDI partner. What should the generator do?