Migration Execution: DMO, Classical, and Beyond
Deep dive into executing SAP migrations with DMO and SUM, classical homogeneous migration, heterogeneous system copy with SWPM, export/import approaches, cutover planning, and selecting the right tool for each scenario.
From strategy to execution
In Module 3, you learned when to use DMO, classical, or heterogeneous approaches. Now we get into the hands-on details of each method β how they actually work, what tools are involved, and what to watch for during execution.
ποΈ Raj rolls up his sleeves. βAlright, we have assessed the landscape, chosen our strategy. Now tell me exactly how the migration runs.β
βοΈ Mei opens the SAP Software Logistics documentation. βLetβs walk through each method step by step.β
DMO with SUM: The replatform workhorse
The Database Migration Option (DMO) runs inside the Software Update Manager (SUM). Think of SUM as the Swiss Army knife for SAP system changes β it handles patches, upgrades, and with the DMO option, database migrations.
Think of DMO like a car engine swap performed while also upgrading the transmission.
Instead of doing two separate garage visits β one to swap the engine (database migration) and another to upgrade the transmission (SAP version upgrade) β DMO does both in a single visit. Your car goes in as an ECC on Oracle and comes out as S/4HANA on HANA. The mechanic (SUM) coordinates both changes so everything works together when you turn the key.
DMO execution phases
Phase 1: Preparation (can be online)
- SUM reads the source system configuration and SAP upgrade stack
- Extracts repository objects and prepares the target schema on HANA
- Imports software components that can be processed without downtime
- The source system remains available to users during this phase
Phase 2: Downtime execution
- Source system is locked β users are disconnected
- Remaining data is migrated from the source database to HANA
- SAP upgrade transformations are applied
- This is the actual downtime window β duration depends on database size and optimization settings
Phase 3: Post-processing (system back online)
- SUM finalizes database indexes, adjusts runtime objects
- System verification checks run
- The system restarts on HANA with the new SAP version
DMO with System Move
When using DMO with System Move to migrate to Azure:
- The source system runs on-premises with the original database
- The target HANA database runs on an Azure VM
- SUM coordinates data transfer between the two environments
- A reliable, high-bandwidth network connection between on-premises and Azure is critical (ExpressRoute recommended)
- After completion, the Azure VM becomes the new production system
Exam tip: DMO prerequisites
DMO requires specific minimum SAP kernel and SUM versions on the source system. The source system must be at a certain patch level before DMO can run. If an exam question mentions an old, unpatched source system, the answer may involve first applying prerequisite patches before DMO can proceed.
Optimizing DMO downtime
For large databases, the downtime phase can be significant. Optimization techniques include:
- Table splitting β Large tables are divided into chunks that migrate in parallel, reducing elapsed time
- Parallel migration streams β Multiple tables migrate simultaneously using multiple worker processes
- Near-zero downtime (nZDT) β An advanced SUM option that uses application-level replication to minimize the conversion window to minutes instead of hours
- Pre-migration of read-only tables β Historical data tables that will not change can be migrated during the preparation phase while the system is still online
ποΈ Raj does a quick calculation. βOur 2 TB database with four parallel streams β what kind of downtime are we looking at?β
βοΈ Mei estimates. βWith good optimization, probably 8-12 hours for a 2 TB conversion. The near-zero downtime option could bring that down significantly, but it adds complexity and requires more preparation. For your six-hour maintenance window, we should seriously look at HSR-based migration for the infrastructure move, followed by DMO conversion once on Azure.β
Classical migration: Backup, restore, and go
Classical (homogeneous) migration is the simplest path when you are keeping the same SAP version and database platform. You are essentially rebuilding the same system on Azure infrastructure.
Backup and restore workflow
- Prepare the target Azure VM β Deploy the VM with the correct OS, install the database software, configure storage
- Take a full database backup on the source system
- Transfer the backup to Azure β Using AzCopy, ExpressRoute data transfer, or Azure Data Box for very large datasets
- Restore the database on the Azure VM
- Reconfigure SAP β Update host names, adjust profiles, configure the SAP instance to start on the new host
- Validate β Run SAP system checks, verify application functionality
Data transfer options for large databases:
- AzCopy over ExpressRoute β For databases up to a few TB, fast enough if you have good bandwidth
- Azure Data Box β For databases above 10 TB or when network bandwidth is limited. Microsoft ships you a physical storage device, you load the backup, ship it back
- Compressed and incremental β Take a full backup early, then an incremental backup just before cutover to minimize the data transferred during the downtime window
Heterogeneous system copy with SWPM
The SAP Software Provisioning Manager (SWPM) handles heterogeneous system copies β changing the database platform without changing the SAP version.
How it works
- Export β SWPM exports all SAP data from the source database into platform-independent files (R3load format)
- Transfer β Export files are copied to the Azure target system
- Import β SWPM imports the data into the new target database (e.g., HANA)
The critical limitation: this is a full data export and import. For a 5 TB database, every row of every table is exported, transferred, and re-imported. It is thorough but slow.
When SWPM is the right choice
- Database platform change WITHOUT an SAP version upgrade (DMO is not applicable because there is no version upgrade to combine with)
- When DMO prerequisites cannot be met (source system too old or too heavily customized)
- Unicode conversion combined with database migration (SWPM handles Unicode conversion natively)
| Aspect | DMO with SUM | Classical (Backup/Restore) | Heterogeneous SWPM | Export/Import (Greenfield) |
|---|---|---|---|---|
| Primary tool | SUM (Software Update Manager) | Database-native backup tools + AzCopy | SWPM (Software Provisioning Manager) | New SAP installation + LSMW/LTMC/third-party |
| SAP version change | Yes β upgrade included | No β same version | No β same version | Yes β new installation at target version |
| Database change | Yes β migrates to HANA | No β same database | Yes β changes database platform | Yes β new installation chooses target DB |
| Downtime driver | Database size and upgrade complexity | Backup transfer and restore time | Full data export and import | New system build plus data validation |
| Rollback strategy | Restore source system from backup | Source system still exists on-premises | Source system still exists on-premises | Keep legacy system running until validation complete |
| Network requirement | High bandwidth between source and target (ExpressRoute) | Sufficient for backup transfer | Sufficient for export file transfer | Minimal β systems are independent |
Cutover planning
The cutover is the actual migration event β the period when the old system goes down and the new system comes up. Poor cutover planning is the number one cause of migration failures.
Cutover checklist
- Define the downtime window β Get business sign-off on the exact start and end times
- Document every step β Create a minute-by-minute runbook with responsible person, expected duration, and verification checkpoint for each step
- Define rollback triggers β At which point do you abort and fall back to the source system? What are the criteria?
- Rehearse β Run the cutover on a sandbox or QA system first. Measure actual timings.
- Communication plan β Who gets notified at each stage? How do users know when the system is back?
- DNS and connection updates β SAP GUI connection strings, web dispatcher URLs, RFC destinations, printer configurations all need to point to the new system
ποΈ Raj nods vigorously. βI have seen migrations fail because nobody updated the RFC destinations. All the batch jobs kept trying to call the old system.β
βοΈ Mei adds it to the list. βExactly. And that is why rehearsal on a non-production system is non-negotiable. You discover those hidden dependencies before they bite you in production.β
Exam tip: Cutover scenarios
Exam questions often present a cutover scenario with a time constraint and ask you to identify the correct sequence of steps or the appropriate rollback point. Practice thinking through the sequence: stop system, take final backup (rollback point), execute migration, verify, update connections, release to users. The final backup before execution is always the rollback point.
Selecting the right tool: A quick decision guide
When an exam question describes a migration scenario, ask yourself:
- Changing SAP version? β DMO (with SUM) or greenfield export/import
- Changing database platform only? β Heterogeneous SWPM or DMO (if combining with upgrade)
- Same version, same database, new infrastructure? β Classical backup/restore or HSR
- Tight downtime window? β HSR-based approach (covered in Module 6)
- Clean start wanted? β Greenfield installation with selective data migration
Knowledge check
PrecisionSteel needs to convert ECC 6.0 on Oracle to S/4HANA on HANA while simultaneously moving to an Azure VM. Which tool and option should Raj use?
Raj is planning the DMO execution timeline for PrecisionSteel's migration. During a DMO execution, at which phase is the SAP system unavailable to users?
Mei is advising a customer with a 15 TB SAP database and limited network bandwidth to Azure. What is the recommended approach for transferring the database backup?
Summary
You now understand the execution mechanics of each migration method: DMO with SUM for combined upgrade and database migration, classical backup/restore for same-platform moves, heterogeneous SWPM for database platform changes without version upgrades, and the critical importance of cutover planning with rehearsal. The System Move option in DMO enables direct migration to Azure VMs during the conversion.
Next, we explore HANA System Replication β a powerful technique for near-zero-downtime migration to Azure.
π¬ Video coming soon