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

Lesson 3 of 5

What drives downtime

10 min

Listen to this lesson

0:00 / 9:48

Business people ask one question about a conversion: how long will we be down? The honest answer is that SUM's technical downtime is only part of the business downtime. Business downtime runs from the moment users are locked out (ramp-down) to the moment they are allowed back after validation (ramp-up). Contoso's 60-hour window, Friday 18:00 to Monday 06:00, has to hold all of it.

  1. Ramp-downJobs stopped, interfaces drained, users locked, final backups
  2. Technical downtimeSUM Execution: switch, main import, table conversion, data conversion, DMO migration
  3. Technical follow-upPostprocessing, kernel and profile checks, SPAU from transports
  4. Application follow-upFinance data migration and other mandatory steps
  5. ValidationTechnical checks, reconciliation, business smoke tests
  6. Ramp-upJobs released, interfaces reopened, users unlocked
Business downtime is the whole chain. SUM's downtime is the middle of it.
The Sign-Off Table · thesignofftable.com/academy/sap-bas-301

The big drivers

DriverWhy it costs timeLevers
Data volumeTable conversions and DMO migration scale with the rows that must be read and writtenArchive and delete before conversion; SAP's Data Volume Management tools help find candidates
Database migration (DMO)Migration of application data is usually the dominant part of a DMO downtimeProcess tuning, table splitting, downtime-optimized DMO (covered in sap-bas-302)
Application data conversionNew S/4HANA data structures (for example in finance and inventory) must be filledClean inconsistencies early; some data can migrate after downtime through SDMI
Manual follow-up stepsFinance migration steps and other activities wait for peopleRehearse, script and parallelize; pre-stage transports and checklists
Hardware and parallelismToo few processes leave resources idle; too many cause contentionMeasure and tune per run; SAP Note 1616401 on parallelism

Typical downtime drivers in a conversion. SAP Note 2351294 collects measures to reduce technical downtime.

One driver is easy to miss: Silent Data Migration (SDMI). Some application data migrations run after the upgrade, during business operations, as background jobs started when the system leaves maintenance mode (monitor them in SDM_MON). SUM can be told to force SDMI processing into downtime so data no longer changes afterward, but SAP warns this can extend downtime significantly. Either way, every SDMI must finish in every client before the next upgrade.

Silent Data Migration runs after downtime by default; forcing it into downtime freezes the data sooner but can lengthen the window considerably.
The Sign-Off Table · thesignofftable.com/academy/sap-bas-301

Downtime-minimizing approaches

Standard mode, tunedShadow system, measured parallelism, data volume reduction. The baseline every project needs.
Downtime-optimized DMOMigrates selected large tables during uptime with record and replay; only when the source is not on SAP HANA.
Downtime-optimized conversionMoves part of the data conversion into uptime; SAP requires an expert trained in ADM329 for Private Edition.
nZDT / ZDO servicesNear-zero downtime service or the Zero Downtime Option for upgrades; expert-led, with extra testing.
Options to shrink downtime, roughly ordered by effort. Each one changes what has to be tested.
The Sign-Off Table · thesignofftable.com/academy/sap-bas-301

Every advanced option has a testing cost. Record and replay means user changes made during uptime are captured and replayed later, so the business has to test that changes made during uptime arrive correctly. The Zero Downtime Option of SUM (SAP Note 2707731) applies to updates and upgrades of systems already on SAP S/4HANA, not to the ECC conversion; under RISE it appears as an Additional Service. Choose the simplest option that fits the window, and decide using measured runtimes, not hope.

Which downtime-minimizing options even apply: ZDO is for systems already on SAP S/4HANA, and downtime-optimized DMO only for sources not on SAP HANA.
The Sign-Off Table · thesignofftable.com/academy/sap-bas-301
  1. Fri 18:00Ramp-downInterfaces drained, jobs suspended, users locked, final checks (5 h in Mock 2)
  2. Fri 23:00SUM technical downtimeSwitch, DMO migration, data conversion (24 h in Mock 2); go/no-go 1 at Sat 18:00 mid-run
  3. Sat 23:00Follow-upPostprocessing checks, finance migration, SPAU from transports (11 h in Mock 2)
  4. Sun 10:00ValidationDan's reconciliations first, smoke tests, go/no-go 2 at 14:00, business validation (13 h)
  5. Sun 18:00Latest fallback startMonday 06:00 minus the 9 h fallback measured in Mock 2, minus a 3 h margin
  6. Sun 23:00Ramp-upJobs released, interfaces reopened, users back by 06:00 (5.5 h in Mock 2)
Contoso's 60-hour window on the cutover clock, built from Mock 2's measured durations (58.5 h, 1.5 h of float). Tuning before the dress rehearsal raised the contingency to the steering group's 4 h minimum.
The Sign-Off Table · thesignofftable.com/academy/sap-bas-301

What would you do?

Contoso's first mock conversion shows 29 hours of SUM technical downtime, and post-processing plus validation took another 32 hours. Grace asks Sam whether the 60-hour window is safe. What is the best answer?

Key points
  • Business downtime runs from ramp-down to ramp-up; SUM technical downtime is one part.
  • Data volume, DMO migration, application data conversion, manual follow-up steps and parallelism drive the time.
  • SDMI moves some data migration after downtime; forcing it into downtime can lengthen the window.
  • Downtime-optimized options exist but add complexity and testing; ZDO is not for ERP-to-S/4HANA conversions.

Check your understanding · not graded

Lesson check

  1. 1. What is the difference between technical downtime and business downtime?

  2. 2. What does Silent Data Migration (SDMI) do?

  3. 3. Which approach does SAP say requires an ADM329-trained expert for SAP S/4HANA Cloud Private Edition?

  4. 4. Which lever reduces downtime before the project even starts tuning SUM?