Should you archive SAP ECC data before migrating to S/4HANA?

Every year of history you carry is converted, reconciled and held in memory. Decide what earns the trip before the runbook is written.

By Chris Benson — 30 years in SAP SD6 min read

The short answer

Yes, for most long-running ECC estates archiving before the migration is one of the cheapest ways to reduce risk, because data volume drives three things you pay for: the conversion downtime window, the reconciliation effort afterwards, and the memory footprint of the HANA system you size and license. Start by measuring which tables are actually large, using DB02 or the DBA Cockpit for table sizes and TAANA to see how that volume splits by year, company code or document type. Then separate the two kinds of reduction: housekeeping of technical data such as application logs, IDocs, job logs and table change logs, which can usually be deleted under an agreed policy, and archiving of business documents through SARA archiving objects, which is constrained by legal retention and audit access. Do the technical clean-up early, archive closed business documents where retention rules allow, and make the archive access path part of the test plan rather than an afterthought.

Why does data volume matter so much in an S/4HANA move?

In a system conversion, data is not just copied. Finance and controlling postings are migrated into the Universal Journal (ACDOCA), material documents into MATDOC, and customers and vendors into Business Partner. Each of those steps runs over the history you bring, and each produces reconciliation work that someone has to sign off. Fifteen years of open and closed documents means fifteen years of conversion runtime, and the business downtime window is set by the longest of those runs plus the checks that follow them.

The second cost is ongoing. SAP HANA is an in-memory database, and the system is sized largely on the data it has to hold. The SAP sizing report for S/4HANA (/SDF/HDB_SIZING, described in SAP Note 1872170) will tell you what your current volume implies, and that number feeds infrastructure cost whether you run on premise, on a hyperscaler or under RISE with SAP. Data you never needed to bring is data you pay to keep warm every month.

How do you find out what is actually taking up space?

Measure before you plan. Table-size views in DB02 or the DBA Cockpit give you the largest tables, and in most aged estates a short list dominates: application logs, IDoc data, change documents, workflow items, and the high-volume document tables in SD, MM and FI. TAANA then breaks the volume of a table down by field values, so you can see how much of a sales document table belongs to closed years, to a company code that was divested, or to a document type nobody uses any more.

That analysis is read-only and it is exactly the kind of work an agent does well against a live system: reading sizes and distributions, mapping each large table to the business process it serves, and showing which share of the volume is closed history. It produces evidence a Basis lead can check, which is what turns archiving from a good intention into a scoped work package.

What is the difference between housekeeping and archiving?

They are different decisions with different owners, and mixing them slows both down. Housekeeping removes technical data that has no business or legal value once it has aged out: it can usually be done by Basis under an agreed retention policy, and much of it can start this month. Archiving moves business documents out of the online database into archive files through SARA and SAP archiving objects, keeping them readable for audit and legal purposes. It needs process owners, finance and often legal to agree residence and retention times per object.

  • Housekeeping candidates: application logs (SLG2 deletion), job logs (RSBTCDEL2), processed IDocs, table change logs, and spool data, each under a written retention policy.
  • Archiving candidates: closed sales orders (SD_VBAK), deliveries (RV_LIKP), billing documents (SD_VBRK), purchasing documents (MM_EKKO), material documents (MM_MATBEL), FI documents (FI_DOCUMNT) and completed work items (WORKITEM).
  • Usually leave alone: open items, documents in an open fiscal year, and anything referenced by a document you are keeping online.

What are the risks of archiving too late or too aggressively?

Archiving has its own lead time. Objects have dependencies, so billing documents cannot be archived while their sales orders and deliveries are still incomplete, and FI documents cannot go while open items remain. Residence-time settings, test runs in a copy of production and sign-off from auditors take weeks, not days. Started three months before cutover, archiving becomes a parallel project competing for the same Basis team that is rehearsing the conversion. Started in the opening weeks of the program, it quietly shrinks every rehearsal that follows.

The opposite risk is archiving data the business or the auditors still reach for. Before you archive an object, confirm how the archived data will be read after the move: which reports and display transactions use the archive, how that works on your target S/4HANA release, and whether a retention or e-discovery requirement calls for SAP Information Lifecycle Management or another archive store. Check SAP's notes for each object on your target release rather than assuming ECC behavior carries over. Then put an archive read into the acceptance criteria, so the first time someone opens a seven-year-old invoice is in a test cycle, not an audit.

How does archiving fit with the rest of the migration plan?

Treat it the same way you treat custom code: retire what does not earn its place, and prove the decision with evidence from the system. Custom code you retire is code you never remediate or regression test; data you archive is data you never convert or reconcile. Both reduce scope, both shorten cutover, and both are cheaper to decide at the start of the program than in the middle of it.

On greenfield or selective data transition paths the question changes shape, because you choose what to load rather than what to leave behind, but the measurement is the same. Know which history the business genuinely needs online, agree it in writing, and let that list drive the load scope, the sizing and the test plan.

Frequently asked

How much downtime does archiving actually save?

There is no reliable rule of thumb, because the saving depends on which tables shrink and whether they sit on the critical path of the conversion. The defensible way to answer is to run a conversion rehearsal on a production-sized copy, archive, and run it again. The measured difference between rehearsals is the number to put in front of the steering committee.

Can we archive after the migration instead?

You can, and some housekeeping is easier afterwards, but you lose most of the benefit. The downtime, conversion runtime and reconciliation effort are all incurred on the data you carry, and the initial HANA sizing is based on it. Post-migration archiving still reduces running cost, so it belongs in the operating model either way.

Who should own the archiving decision?

Basis owns the mechanics, but the retention decisions belong to the business: finance and the process owners for each document type, with legal or internal audit confirming retention periods. A single named coordinator on the customer side keeps those decisions moving, because archiving stalls whenever every object needs a separate committee.

See your own Clean Core Score.

Run the free S/4 readiness scan in under a minute — no system access required.

Keep reading