On the day an ECC system becomes an S/4HANA system, most of the visible work belongs to the business: testing, sign-off, training, a new login screen. Behind that sits the Basis team, which has to make the technical conversion run, repeat, finish inside an agreed window and, if it goes wrong, put everything back. Basis work rarely gets a line in the business case. It decides whether the go-live weekend goes to plan.
This guide explains what the Basis team does in a system conversion, in plain English. It covers the tools, the landscape, the Software Update Manager (SUM) and the Database Migration Option (DMO), test conversions, cutover and the weeks of support after go-live. It is written for programme managers, IT leaders and business sponsors who need to understand and challenge the technical plan, and for Basis engineers who want a clear checklist.
What the Basis team owns in a conversion
Basis is the technical administration layer of SAP: installing and updating systems, managing the database and operating system relationship, handling transports, monitoring performance and keeping the landscape healthy. In a conversion that role widens, because the conversion is largely a systems operation.
The Basis team typically owns these responsibilities, even where other teams contribute:
Technical readiness: confirming that the source system meets prerequisites such as Unicode and a single ABAP stack, and that operating system and database versions are supported for the target.
Landscape design: deciding which systems are converted in which order and how development, quality and production stay aligned while they differ.
Tooling: generating the stack file with the Maintenance Planner, downloading the right SUM version and software, and preparing the target database.
Conversion runs: executing SUM, including DMO where the database changes, in the sandbox, in test systems and in production.
Measuring and tuning: recording how long each phase takes and reducing it where the plan needs a shorter window.
Cutover execution: running the minute-by-minute plan on go-live weekend with the other teams.
Fallback: owning the backups, the restore tests and the technical conditions for going back.
Hypercare: monitoring, performance tuning, batch job and interface checks, and fast response in the first weeks.
Bring the Basis lead into the route decision, not just into the execution. The choice between a standard conversion, a downtime-optimized conversion and a different route has consequences for downtime, planning effort and risk that the Basis team can estimate better than anyone.
The tools: SUM, DMO, Maintenance Planner and doC
A small set of SAP tools does most of the technical work. The table summarises them in business-friendly terms.
| Tool | What it does | Why it matters to the plan |
|---|---|---|
| Maintenance Planner | SAP-hosted tool that checks that add-ons and products are compatible with the target release and generates the stack file used as input for SUM. | A blocked add-on stops the plan early. Run it first. |
| Software Update Manager (SUM) | The tool that carries out the conversion. It updates the software, converts the data and switches the system to the new release. | Its runtime defines most of the downtime. |
| Database Migration Option (DMO) | A feature of SUM that combines the software update with the migration to SAP HANA in one procedure, with a single downtime. Enabled automatically when the source database is not SAP HANA. | Lets you change database and release in one step, with a built-in fallback. |
| DMO with system move | Variant of DMO that moves the system to a different environment during the conversion. | Useful when changing hosting or hardware at the same time. |
| Downtime-optimized conversion (doC) | An approach within SUM that moves parts of the data conversion into uptime and replays user changes later, to shorten downtime. | Helps with tight windows, but adds planning and test effort. |
| SPDD, SPAU, SPAU_ENH | Adjustment transactions for dictionary objects, repository objects and enhancements that were modified. | Work for the ABAP team, scheduled by Basis around the runs. |
| SODH (Obsolete Data Handling) | Deletes obsolete data left after conversion, once the system is validated. | Frees space after sign-off. Not a SUM phase. |
SAP's community documentation states that all conversions from ERP to S/4HANA must use SUM 2.0 or later, and that you should download the latest support package of SUM from SAP's software logistics toolset page. Tool versions change often, so check SAP's current conversion guide for your target release before you plan.
Landscape and preparation
Check the starting point
Before any tool runs, confirm the prerequisites. The source needs to be SAP ECC 6.0 at a supported enhancement package level, Unicode, and ABAP single stack. A dual-stack system, with both ABAP and Java, has to be split first. For DMO, the source and target operating system and database must be supported, which you verify in SAP's Product Availability Matrix for the source and target releases. The target SAP HANA database needs to be installed and configured, and sized for what it will hold after conversion.
Generate the stack file
Use the Maintenance Planner to plan the transition. It checks add-ons and products, flags third-party add-ons with unknown status for confirmation with the vendor, and, if everything is compatible, produces the stack file that SUM uses. Add-ons that are no longer needed must be uninstalled during the conversion, and the vendor must supply the uninstaller. If none exists, it can block the conversion. Raise these early.
Design the landscape sequence
A typical landscape has development, quality assurance and production systems, plus sandboxes. A common pattern converts a sandbox copy first, then the development and quality systems, then production. The difficulty is that for a period, systems on different releases sit in the same transport route. Many programmes manage this by freezing changes in the old landscape, running a separate maintenance track for urgent fixes, and then retrofitting those fixes. Basis owns the transport rules that make that work, and the programme manager owns the freeze dates.
Prepare data and housekeeping
Every gigabyte you do not convert saves time. Archive or delete what the business agrees it does not need. Clean up old spool requests, logs and temporary data. Make sure customer and vendor data is ready for the business partner conversion, which is a functional task that Basis may need to coordinate.
Agree the hand-offs with the other teams
A conversion fails at the seams between teams more often than inside any one of them. Basis needs specific inputs from others and owes specific outputs in return. Settle them in writing during preparation.
| Team | Basis needs from them | Basis provides |
|---|---|---|
| ABAP and development | Adjusted objects, transports ready for each run, agreed freeze dates. | Converted systems on schedule, SPDD and SPAU windows, transport route. |
| Functional and business teams | Test calendars, validation owners, sign-off criteria. | A fresh converted copy for each test cycle, with the date and the known issues. |
| Security and authorisations | Role changes for the new release and Fiori. | A system where roles can be loaded and tested before go-live. |
| Integration | Interface inventory, endpoint changes, partner contacts. | Connection and certificate set-up, switch-over steps in the runbook. |
| Infrastructure and hosting | Hardware for HANA, storage, network, backup capacity. | Sizing, performance results and capacity forecasts. |
Plan backups and sizing
Backups are your fallback. Schedule full backups of the source system before each run, and test restores. Size the target for the converted database, any shadow system needs and the space for export or migration files if you use DMO with system move. SAP Readiness Check includes a sizing section that gives an early estimate; repeat the exercise after archiving.
How SUM runs a conversion: uptime and downtime
SUM is built so that as much work as possible happens while the source system is still in production. This is the idea behind the uptime and downtime split, and understanding it is the key to understanding a conversion window.
Uptime: the shadow system
In SUM's standard mode, a second instance, called the shadow instance, runs on the same database using the target release's kernel. It cannot access your application data. While production carries on, SUM builds the new release's repository in shadow tables, imports the dictionary objects, activates them and distributes them. This part can take hours, because for example activating large numbers of dictionary objects takes time, but it happens in uptime and does not affect users. SAP's documentation notes that for target releases from S/4HANA 1909, the shadow repository can be created on the target database, and for earlier targets it sits on the source database.
Adjustments to modified objects, using SPDD, are handled around this phase. The shadow system also supports batch monitoring and troubleshooting.
Downtime: the switch and the conversion
Downtime begins when users are locked out and the profile changes are made. In the standard approach, the shadow instance is stopped before downtime starts. SUM then switches to the new kernel and the new repository, converts the application tables, imports remaining upgrade data and language data, and runs the post-processing programs. For an S/4HANA conversion, the step that runs the XPRA and after-import programs is where the actual data conversion happens, such as the move to the new finance data model. SAP's documentation also notes that in standard mode, downtime does not depend on the number of languages, support packages or add-ons.
DMO: changing the database in the same run
When the source database is not SAP HANA, SUM uses DMO to migrate during the same procedure. The shadow repository and tables are prepared in uptime. At the start of downtime, the connection switches to the target database, and the application data is migrated and converted. R3load processes move data in parallel. SAP highlights two safeguards. The source database is not changed, so it remains available as a fallback. And SUM includes a reset function that can take the procedure back at defined points. Data can move through a pipe, without intermediate files, which speeds up the transfer in standard DMO.
Downtime-optimized conversion
Where the standard window is too long, doC moves more work into uptime. Part of the financial data migration, field conversions for some large tables and the migration of large application tables that are not affected by the new data model can run while users are still working. SUM uses database triggers to record changes that users make in the meantime and replays them during downtime. SAP is clear about the cost: the approach needs an expert trained on the method, more planning and test effort, restrictions on production in the period before downtime such as a freeze on customising and development, and several runs on production copies.
doC is not a free speed-up. Every freeze and restriction it requires lands on the business, and replaying changes adds risk. Choose it when a measured standard run cannot fit the window, not because a shorter number looks better in a presentation.
Choosing between standard and downtime-optimized
The choice is best made on evidence from your own test conversions. The decision table below shows the questions to ask, and who owns each answer.
| Question | If yes | If no | Owner |
|---|---|---|---|
| Does a measured standard run fit the agreed downtime window with margin? | Stay with the standard approach. | Look at tuning first, then doC. | Basis lead |
| Can the business accept a freeze on customising and development before cutover? | doC is possible. | doC is not suitable. | Programme manager, process owners |
| Can production tolerate the restrictions on mass loads during replication? | doC is possible. | Schedule around them or avoid doC. | Operations, Basis |
| Do we have, or can we get, an expert trained in the method? | Plan the extra test runs. | Avoid doC or bring one in. | CIO, Basis lead |
| Is there time for several extra runs on production copies? | doC is realistic. | The risk outweighs the gain. | Programme manager |
If the answers point to the standard approach, spend the saved effort on tuning the standard run instead. Parallelism settings, hardware for the migration, table clean-up and archiving before conversion all shorten a standard run without extra process risk. The best downtime reduction is often the data you did not need to convert in the first place.
After the conversion
Once the system is technically converted, functional teams validate and the ABAP team completes adaptations. After successful validation, obsolete data can be removed with the Obsolete Data Handling transaction, SODH. Basis also re-establishes everything around the system: kernel parameters, the transport route, monitoring, backups, printers, batch schedules, interfaces and the Fiori front-end configuration where relevant.
Test conversions and rehearsals
A conversion that has run once has not been tested. It has been tried. Plan several runs, each with a purpose.
| Run | System | What it proves | Basis outputs |
|---|---|---|---|
| First sandbox conversion | Copy of production, often reduced or with scrambled data | That the tools run end to end; what errors appear; which prerequisites were missed. | Issue log, first timings, list of fixes. |
| Development system conversion | Development | That the code base converts and developers can work on the new release. | Adjusted objects, transport route plan. |
| Test conversions on production copies | Full copy of production | Realistic runtime and data behaviour; input to business testing. | Timed phase log, tuned parameters, a refined runbook. |
| Dress rehearsal | Full copy of production | The complete cutover plan, with all teams, in sequence and on the clock. | A final timeline, go and no-go criteria, confirmed fallback. |
| Production cutover | Production | The real run. | Executed runbook, post-conversion evidence. |
The most important rule is to time everything. Record the duration of every phase in every run, then look at the shape. Which phases dominate? What changed between runs? A run that is faster than the last one because someone tuned the database is useful knowledge. A run that is faster because the data was smaller than production is a trap.
Test conversions on full copies of production also feed business testing. Finance can reconcile balances after the data conversion. Supply chain can run end-to-end process tests. Defects found there come back to the technical team. Build the rehearsal calendar so that each business test cycle has a fresh, converted copy to work on.
Keep one runbook and one decision log from the first sandbox run onwards. By the dress rehearsal, the runbook should be good enough that someone who was not on the project could run it. If it is not, the single point of dependency is a person, not a plan.
Cutover: go-live weekend
Cutover is the controlled sequence that turns the old system into the new one. For Basis it covers the technical conversion. For the programme it also includes business freeze, data checks, interface switch-over and the go and no-go decision.
What a cutover plan contains
A minute-by-minute schedule, with an owner for each task and expected duration based on measured rehearsals.
Pre-cutover tasks: final backups, freeze confirmations, pre-checks, user lock, queue and job clean-down.
Technical conversion steps: SUM downtime phases and any migration, with monitoring points.
Post-conversion tasks: kernel and parameter settings, interface activation, job reactivation, authorisation checks and sanity tests.
Business validation steps and their owners, with the people actually available on the day.
Go and no-go checkpoints, each with explicit criteria and a named decision-maker.
The fallback plan, with the latest point at which it is still possible.
A communication plan, so everyone knows who is talking to whom and when.
A worked example (illustrative)
The timeline below is invented to show the structure of a cutover. The durations are not benchmarks, and a real plan uses your own rehearsal timings.
The situation. A company has agreed a cutover window from Friday evening to Sunday evening with the business. Rehearsals have shown a technical downtime of about a day when the plan runs well. The team has built in margin before the business validation time on Sunday.
| When | Step | Owner | Checkpoint |
|---|---|---|---|
| Friday 18:00 | Business freeze, final transports imported, jobs paused, users locked | Programme manager, Basis | Go or no-go 1: freeze confirmed |
| Friday 19:00 | Full backup of source system and restore point confirmed | Basis | Backup verified |
| Friday 20:00 | SUM downtime starts: switch and data conversion | Basis | Timing against rehearsal logged hourly |
| Saturday 12:00 (latest) | Technical conversion complete | Basis | Go or no-go 2: fallback still possible |
| Saturday 12:00 to 18:00 | Post-conversion tasks: parameters, interfaces, authorisations, jobs, sanity tests | Basis, ABAP, integration | Technical sign-off |
| Saturday 18:00 to Sunday 12:00 | Business validation: balances reconciled, key processes tested | Finance, supply chain, sales | Go or no-go 3: business sign-off |
| Sunday 14:00 | Users released, hypercare starts | Programme manager | Go-live announced |
What the example shows. Three things stand out. There is margin between the expected end of the conversion and the latest point of no return. The fallback decision is made while there is still time to carry it out. And the business validation window is protected, because a conversion that finishes technically but leaves no time to check balances is not finished.
Fallback
The fallback plan needs a decision point, a named decision-maker and a tested method. With DMO, the source database remains unchanged, so returning to it is possible until you start using the new system for real. After business users have posted new transactions in the converted system, going back means losing or re-entering those. Define the point at which you accept that and what you do if you pass it.
Hypercare: the first weeks after go-live
Hypercare is a common industry term for the period of extra support after go-live. It is not an SAP tool. For Basis it is when conversion issues that tests missed show up, usually as performance, jobs and interfaces.
| Topic | What to watch | Typical action |
|---|---|---|
| Performance | Slow transactions and reports, expensive database statements. | Use the SQL Monitor on production to find costly statements; hand findings to ABAP for optimisation. |
| Batch jobs | Jobs that fail, run late or run twice. | Compare to the pre-conversion schedule; fix dependencies and variants. |
| Interfaces | Failing or delayed messages to and from other systems. | Monitor queues; confirm endpoints and certificates; work with the integration lead. |
| Authorisations | Users missing access to Fiori apps or transactions. | Work with security on role changes; fix missing authorisations quickly. |
| Database and memory | Memory growth, backup duration, log space. | Review sizing against actual use; adjust backups and housekeeping. |
| Transports and development | Changes held back during the freeze; retrofits of urgent fixes. | Reopen the transport route; retrofit fixes from the maintenance track. |
| Clean-up | Obsolete data and temporary objects left after conversion. | After validation, run obsolete data handling; remove temporary setup. |
Agree the length and staffing of hypercare before go-live. It needs a rota, a triage process that separates technical from functional problems, a daily review in the first week, and an exit criterion so it does not drift on indefinitely. Many programmes then transition to a normal support model with defined service levels. A partner that knows the conversion, because it ran it, is well placed to carry on supporting the system.
Risks and watch-outs
Planning from estimates instead of measurements
Downtime estimates from slide decks and from other companies' systems do not transfer. Your table sizes, hardware and data are your own. Only timed runs on a copy of your production data give a number you can promise.
Rehearsing on unrepresentative data
A reduced sandbox is useful for learning the tools. It says little about production runtime. Use a full copy of production for the timing runs.
Ignoring the change freeze
The longer the project, the more change piles up in the old landscape. Without a clear freeze and a maintenance track, test systems drift from production and the conversion meets surprises.
Underestimating the human side of cutover
A runbook is only as good as the people running it. Cutover weekends fail through tiredness and unclear ownership as often as through technical faults. Plan shifts, handovers and rest.
No decision on fallback
If the fallback has never been tested, it is a hope. Restore the backup in a rehearsal at least once and time it.
Assuming the end of the conversion is the end of the work
The system will be slower in places and quieter in others than before. Performance tuning, batch re-scheduling and authorisation fixes are normal. Budget for them.
Do not let the same few people carry every run, every fix and the cutover. A conversion that depends on one expert's memory is fragile. Pair people, write down what you learn after each run, and rotate the lead role between rehearsals.
Basis checklist for a system conversion
- Source system confirmed as ECC 6.0 at a supported level, Unicode and ABAP single stack, or a plan exists to get there.
- Operating system and database versions checked against SAP's Product Availability Matrix for source and target.
- Maintenance Planner run, add-ons confirmed or uninstall plans agreed, and the stack file generated.
- Latest SUM version downloaded and the target SAP HANA database installed, configured and sized.
- Landscape sequence, change freeze and maintenance track agreed with the programme manager.
- Archiving and housekeeping completed, with the agreed data left out of the conversion.
- Full backups scheduled before every run, and a restore tested.
- First sandbox conversion completed, with the issue log and timings recorded.
- Test conversions run on full production copies, with every phase timed.
- A decision made on standard conversion or downtime-optimized conversion, based on measured runs.
- One runbook maintained from the first run, with owners and expected durations.
- Dress rehearsal completed on the clock, with go and no-go criteria agreed.
- Fallback plan documented, tested and approved, with the latest decision point defined.
- Hypercare rota, triage process and exit criteria agreed before go-live.
- Obsolete data handling and clean-up scheduled after business validation.
How Steady Rabbit can help
Steady Rabbit's Basis and Support team plans and runs conversion technical work, from the Maintenance Planner and landscape design through test conversions and cutover to hypercare and ongoing support. The engineers work alongside our ABAP specialists, which keeps custom code adaptation and technical conversion on one timeline. We are India-based, have delivered SAP work since 2016, and have about 60 engineers across the company, with team leads who bring 7 to 16 or more years of SAP delivery and certifications in ABAP Cloud and SAP BTP. We work directly with enterprises and as a subcontract partner inside SAP partners' programmes.
Talk to a Basis lead Basis and Support S/4HANA Migration ABAP Code Modernization Services For SAP partners
FAQs
What does the Basis team do in an S/4HANA conversion?
Basis prepares the technical landscape, generates the stack file with the Maintenance Planner, runs the Software Update Manager and, where needed, the Database Migration Option, times and repeats test conversions, executes cutover, owns the fallback and supports the system in hypercare.
What is SUM?
SUM is SAP's Software Update Manager, the tool that performs the conversion. It updates the software, converts the data to the new model and switches the system to the new release, with most of the preparation running while the source system stays in production.
What is DMO and when is it used?
The Database Migration Option is a feature of SUM that migrates the database to SAP HANA as part of the same procedure as the update or conversion. It is enabled automatically when the source database is not SAP HANA, and it keeps the source database unchanged as a fallback.
How long will the downtime be?
It depends on system size, hardware, data and the approach, so no responsible figure can be given without measured test conversions on a copy of your production system. Plan to measure it and to design the window around the result.
What is downtime-optimized conversion?
It is an approach within SUM that moves parts of the data conversion into uptime and replays user changes during downtime. It can shorten the downtime but adds planning and testing effort, needs a trained expert, and places restrictions on production beforehand.
How many test conversions should we run?
SAP's documentation advises several runs on production copies when using downtime-optimized conversion, and most programmes run several in any case. Plan for a sandbox run, development and quality conversions, multiple production-copy runs and a dress rehearsal.
Can we go back if the conversion fails?
With DMO the source database stays unchanged and the procedure can be reset, so a return is possible up to a defined point. Define that point, test the restore in rehearsal and agree who makes the call.
What is hypercare?
Hypercare is the period of intensified support after go-live, usually a few weeks, in which the team monitors performance, jobs, interfaces and authorisations, and fixes issues that testing missed. It ends against agreed exit criteria.
Can Steady Rabbit run the Basis work for our conversion?
Yes. We can plan and run the technical conversion, or supply Basis engineers into your partner's programme. Use the contact form on this page to start.
Sources
SAP Support: Database Migration Option (DMO) of Software Update Manager. DMO scenarios, shadow repository, fallback through reset, downtime-optimized DMO.
SAP Support: Downtime-optimized conversion approach of Software Update Manager. What doC moves into uptime, its prerequisites and limitations.
SAP Learning: SAP S/4HANA conversion and SAP system upgrade, Processing the SUM. SUM phase names, shadow system, uptime and downtime phases, obsolete data handling.
SAP PRESS Blog: SAP S/4HANA conversion, what is the Software Update Manager? SUM 2.0 requirement, standard and downtime-optimized modes, shadow instance, DMO.
SAP Community: Handling add-ons during a system conversion or upgrade to SAP S/4HANA. Maintenance Planner, stack file and add-on uninstallers.
SAP Community: How to move to SAP S/4HANA. Prerequisites for a system conversion (ECC 6.0, Unicode, ABAP single stack).
SAP Community: S/4HANA system conversion, custom code adaptation process. SPDD, SPAU and SPAU_ENH adjustments, SQL Monitor and the post-conversion phase.
SUM versions, note numbers and tool behaviour change between releases. We checked these sources in November 2026. Confirm current details in SAP's conversion guide for your target release.