FarrenioFarrenio
SAP on AWS

S/4HANA on AWS: brownfield vs greenfield vs selective transition

How to choose a SAP S/4HANA migration path, the questions that actually decide it, and where AWS landing-zone design and HANA-certified sizing fit once the path is settled.

12 min read

"Should we go greenfield or brownfield?" is the question every S/4HANA programme opens with, and it is usually the wrong place to start. The path is a consequence of decisions about your current system, not a decision in itself. This post works through what actually drives the answer, the real tooling behind each route, then where the AWS-side sizing fits once the path is settled.

The three paths

PathWhat it isKeepsCosts you
GreenfieldNew S/4HANA implementationNothing by default; a clean rebuildLargest project; re-test & re-train everything
BrownfieldIn-place system conversion of ECCCustomising, history, most configCarries forward the technical debt you already own
SelectiveNew system, selected config + data moved inWhat you choose: company codes, date rangesMost tooling-dependent; needs a transition partner

The questions that actually decide it

Rather than starting from the path, start from these four. The answers point at the path on their own, and notice that none of them is an AWS question.

QuestionIf the answer is……it points at
How much of today's system is worth keeping?"Mostly scar tissue" / "It broadly works"Greenfield / Brownfield
Appetite for business disruption?High mandate / Low toleranceGreenfield / Brownfield
How clean is the data?Cleansing already happening / Fine as-isGreenfield or Selective / Brownfield
Deadline pressure (ECC maintenance)?Hard date loomingBrownfield (fastest to "running")

Choosing wrongly is expensive in both directions: a greenfield rebuild where a conversion would do burns a year the business didn't have; a brownfield conversion where the requester needed a clean core drags the old mess into the new system.

Run the readiness checks before you commit

The path is not purely a judgement call. SAP ships tooling that turns "how much debt do we carry?" into a report. Run these on the source system before the decision is final:

ToolTells youRoute it informs
SAP Readiness CheckSimplification items, add-on & sizing impact, custom-code footprintBrownfield feasibility
Simplification Item Check (SI checks)Which functional simplifications block a conversion, and the mandatory pre-stepsBrownfield
ATC, S4HANA_READINESS variantCustom ABAP that won't survive the conversion (obsolete tables/APIs)Brownfield custom-code adaptation
SUM with DMOThe actual in-place run: upgrade + DB migration to HANA + conversion, one procedureBrownfield execution
SPDD / SPAUWhere your modifications collide with SAP objects during the conversionBrownfield modification adjustment

A high Readiness Check score with few SI blockers and a small ATC finding count is the evidence that argues for brownfield. A daunting one is the evidence that a clean core, whether greenfield or selective, the latter usually with SAP's Data Management & Landscape Transformation tooling or a partner product, is the cheaper answer in disguise.

Where AWS fits, and where it does not

AWS enters once the path is chosen, and it is largely the same infrastructure problem either way:

  • Sizing the target. S/4HANA needs HANA-certified compute. The certified EC2 families change over time, covering the large memory-optimised (x2idn/x2iedn, r-family) and, for very large HANA, the bare-metal u- family, so size against the current certified list, not a number from a past project.
  • The landing zone underneath. Either path lands in an AWS account structure with a network, security baseline and IAM model. Building that well once is independent of the migration path, and we keep a working checklist for SAP-on-AWS landing zones.
  • The cutover. Brownfield conversions and greenfield go-lives both need a rehearsed window with a rollback plan. Migration Hub and Application Migration Service support both; the wave planning differs.
  • HA and DR. Productive HANA gets HANA System Replication across Availability Zones and a DR posture across regions, regardless of how it got there.

The one place the path changes the AWS work meaningfully is the transition window. A brownfield conversion often runs source and target in parallel for a period; a greenfield build runs the new system alongside the old until cutover. Both need capacity planned for the overlap, and both benefit from non-production auto-shutdown so the parallel run doesn't double the bill for longer than necessary.

And RISE?

RISE with SAP is a fourth option that cuts across this: SAP runs the infrastructure for you, on a hyperscaler, under one contract. It is the right answer for some landscapes and the wrong one for others. The comparison is in RISE with SAP vs self-hosted on AWS. Many mid-size estates end up hybrid: some workloads on RISE, others on a self-hosted AWS landing zone next to it.

A practical sequence

  1. Run the Readiness Check and ATC on the source. Let the reports, not opinions, size the debt.
  2. Answer the four questions with the business, not just IT. The path falls out of the answers plus the reports.
  3. Size the S/4HANA target against the current HANA-certified instance list.
  4. Build the landing zone properly. It outlives the migration and you don't want to rebuild it later.
  5. Plan the cutover and the rollback, then rehearse the cutover before the real one.
  6. Stand up monitoring before go-live, not after. Week one on a new system is exactly when you most want one view of work processes, jobs, dumps and HANA signals.

That last point is where Farrenio fits after the move: SM50, SM37, ST22 and the rest cross-system, alongside HANA database signals, in one view from day one. As an AWS reseller for SAP, the landing zone, the sizing and the operate-after-go-live are the work we do.

If you want a sized S/4HANA-on-AWS landing zone and an honest read on which migration path fits your situation, write to contact@farrenio.com.

Run Farrenio against your own SIDs.

14-day sandbox tenant. No card. Real data.

Book a demo