Overview

This page contains 55+ real SAP S/4HANA AWS migration interview questions and short technical answers covering cloud architecture, SUM DMO migration, HANA HA/DR setup, RISE with SAP, BTP integration, GenAI in SAP, automation, and performance optimization.

The interview guide is based on a real technical round for a Senior Delivery Consultant — SAP Migration role at a major cloud hyperscaler professional services organization.

All company names, candidate names, and private identifiers are fully anonymized.

Use these answers as short, technical, interview-ready responses that demonstrate hands-on ownership and production depth.

Entity Summary

Role: Senior Delivery Consultant — SAP Migration

Skills: SAP S/4HANA, SUM DMO, HANA HSR, AWS EC2, BTP Integration Suite, RISE with SAP, SWPM, Pacemaker, AWS Backint Agent, CloudFormation, Step Functions, Bedrock, Textract, FAISS, UiPath RPA

Domains: Cloud Migration, SAP Architecture, Healthcare IT, Telecom Enterprise, AI/GenAI Integration

Use Cases: Brownfield S/4HANA migration, Greenfield implementation, SDT consolidation, RISE migration, HANA HA/DR, GenAI invoice processing, RPA automation

Interview Type: Deep technical round — Senior/Lead level

Current Technology Context: SAP S/4HANA 2023, RISE with SAP on AWS, BTP Integration Suite, AWS Bedrock GenAI, SUM 2.0 SP17+

Key Topics Covered: Migration methodology, SUM DMO phases, HANA HA setup, DR architecture, SWPM classical migration, tenant copy, backup setup, application server HA, overlay IP, EFS synchronization

Need Real-Time SAP Interview Support?

WhatsApp ProxyTechSupport now: +91 96606 14469
https://wa.me/919660614469

We provide live real-time support during SAP technical interviews, proxy interview assistance, and SAP migration project support. Our AWS job support team covers HANA, BTP integration, and cloud migration scenarios end to end.

How to Use This Guide

These are real questions asked in a live Senior Delivery Consultant SAP Migration interview.

Read each question. Read the short technical answer. Internalize the ownership language — not textbook definitions. Practice speaking each answer in 60 to 90 seconds. Be ready for the follow-up questions listed after each answer.

Interview Questions and Answers

Q1. Tell me about yourself.

I have 15 years of experience leading large-scale SAP and cloud transformation programs — strategy, architecture, and production delivery end to end.

Most recently I led the technology integration of a multi-billion dollar acquisition at a large enterprise telecom company — assessed 535 applications, built the disposition framework, drove the post-merger roadmap, and managed $50M in annual professional services revenue.

Before that I was Director Architect at a healthcare enterprise — led SAP S/4HANA brownfield migration on AWS using SUM DMO, HANA System Replication HA design, and BTP integration. Delivered 60% ERP efficiency improvement and 70% infrastructure cost reduction.

Earlier I built deep SAP functional expertise at global consulting firms — SD, MM, FICO, R2R.

What connects everything is hands-on end-to-end ownership — not just designing on a whiteboard but writing the cutover plan, catching sizing errors before go-live, and standing in front of the CIO with SUM DMO simulation logs explaining delays with data and options.

Follow-up defense:

  • What was the most complex SAP program you delivered?
  • How did you manage a team of that size across an M&A?
  • What specific AWS architecture did you design for SAP?

Q2. Why do you want to move from a Director role to a Senior Delivery Consultant role?

The title is not the reason. The work is.

As Director and Senior Director my week became budget cycles, headcount planning, vendor negotiations, and escalation calls. I moved further from technical delivery.

The Senior Delivery Consultant role puts me directly in front of the customer CTO as the technical authority. I own the architecture decision, the migration execution, the cutover outcome — no management layer between me and delivery.

That is more technically demanding than what I do today as Senior Director.

I am trading organizational management overhead for direct technical impact at customer level. That is a deliberate choice.

Follow-up defense:

  • Will you be comfortable going back to individual contributor work?
  • How do you handle situations where you disagree with the customer?

Q3. Walk me through your SAP migration experience — specifically migration architecture and delivery.

At a healthcare enterprise I owned the full brownfield S/4HANA migration on AWS end to end.

Ran SAP Readiness Check 2.0 against production ECC. Custom Code Migration app on BTP scanned all Z-programs — found 3.2x more custom code than the verbal estimate. That catch saved us from a failed SUM DMO run.

Sized HANA using SAP Quick Sizer and DBACOCKPIT column store analysis. Production landed on x2idn.32xlarge at 2 TB RAM — a 40% sizing error I caught over the initial proposal.

SUM DMO migration execution. First dress rehearsal showed 22-hour runtime. Negotiated 28-hour business downtime window. Cutover ran 18.5 hours — clean go-live.

Delivered 60% ERP process efficiency improvement and 70% infrastructure cost reduction.

Follow-up defense:

  • What specific SUM DMO phases consumed the most time?
  • How did you handle the custom code remediation?
  • What AWS instance types did you select and why?

Q4. Walk me through SAP S/4HANA on AWS architecture end to end.

Landing zone first — VPC with separated subnets for HANA tier, application tier, and DMZ. Transit Gateway for multi-account connectivity. Direct Connect with BGP failover to VPN.

HANA sizing from SAP Quick Sizer SAPS rating and DBACOCKPIT column store analysis. Production typically r6i.32xlarge at 1024 GB or x2idn.32xlarge at 2 TB for analytics-heavy workloads.

Storage — gp3 EBS for HANA data and log on separate attachments — never shared. HANA log volume minimum 7000 IOPS. Analytics-heavy — io2 Block Express at up to 256,000 IOPS.

HA — HANA System Replication synchronous mode across two AZs. Pacemaker overlay IP with stonith fencing via AWS APIs. Route 53 health checks for DNS failover.

Backup — AWS Backint Agent streaming to S3 via HANA backint interface. Configured in HANA global.ini with S3 bucket ARN and KMS key. Full weekly plus 15-minute log backups.

Monitoring — SAP Data Provider pushes HANA memory, dialog response time, work process queue depth to CloudWatch. All admin access via SSM Session Manager — no open SSH ports.

Follow-up defense:

  • Why x2idn over r6i?
  • How do you configure Pacemaker stonith on AWS?
  • What CloudWatch alarms do you configure?

Q5. You integrated AI with SAP — what was the use case?

Invoice exception handling inside SAP at a large enterprise.

When a vendor invoice fails three-way match in S/4HANA — PO, goods receipt, and invoice misalign — it goes to a human reviewer queue. That queue had 400 to 600 exceptions per week.

SAP workflow fires an RFC on exception creation. RFC triggers a BTP iFlow. BTP iFlow pushes the invoice PDF to S3. Lambda calls Textract AnalyzeDocument with TABLES and FORMS feature types — extracts vendor number, line items, totals, tax amount.

Lambda calls Bedrock InvokeModel with Claude — passes Textract output plus vendor master data from DynamoDB plus open PO context from a FAISS vector index on EFS.

Bedrock returns match_confidence score, recommended_action, exception_reason as JSON. Lambda posts back through BTP iFlow to SAP Invoice Management OData API. Human review only when confidence below 0.75.

End-to-end latency — 3 to 6 seconds per invoice. Human review queue dropped 65%.

Follow-up defense:

  • How did raw SAP data get into the LLM?
  • What was the FAISS index structure?
  • How did you handle hallucinations?

Q6. How does raw SAP data become LLM-consumable?

Raw SAP data never touches the LLM directly. Lambda is always the transformation layer.

SAP stores data across normalized tables — EKKO, EKPO, MKPF, MSEG. You cannot pass raw table rows to an LLM.

First — extract via OData V4 API, not direct table access. Purchase Order API, Goods Receipt API, Business Partner API. Response comes back as structured JSON with business field names.

Second — Lambda flattens the nested JSON into a clean structured object. One record per line item — vendor_number, po_number, ordered_quantity, unit_price, confirmed_gr_quantity.

Third — Textract extracts the invoice side — billed_quantity, unit_price, totals.

Fourth — Lambda assembles a prompt context block combining PO from SAP OData, invoice from Textract, vendor master from FAISS vector index. Claude sees business-readable structured context — not EKKO or EKPO.

Fifth — Bedrock returns structured JSON. Lambda posts to SAP via BTP iFlow OData call.

Follow-up defense:

  • How did you handle large PO line item sets exceeding context window?
  • What chunking strategy did you use for FAISS?

Q7. How does an email attachment access the processing pipeline?

Email arrives in monitored mailbox — Exchange or Microsoft 365.

Option 1 — Microsoft Graph API. Lambda polls mailbox, retrieves PDF attachment as base64, writes to S3 with structured key. Marks email as read.

Option 2 — Amazon SES receipt rule. Vendor email routes through SES — SES rule writes raw email including attachment directly to S3. Lambda parses MIME message, extracts PDF, writes to clean S3 prefix.

Option 3 — SAP Business Network. Invoice arrives as cXML — PDF embedded as base64 in cXML. BTP iFlow extracts base64, writes to S3.

From S3 onward the pipeline is identical regardless of ingestion path — S3 event triggers Lambda, Lambda calls Textract, Lambda calls Bedrock, BTP iFlow posts back to SAP.

Metadata tagging on S3 — vendor_email, received_timestamp, po_number, processing_status. After completion — status updated to complete with match_confidence. Full audit trail.

At the large enterprise implementation end to end from email arrival to SAP exception update was under 10 seconds.

Follow-up defense:

  • How did you handle malformed or password-protected PDFs?
  • What error handling did Lambda have?

Q8. What are the prerequisites to execute SUM DMO migration from ECC to S/4HANA?

Four discovery tools must run before any migration recommendation.

SAP Readiness Check 2.0 — runs against production ECC. Outputs custom code impact, simplification item list, open business process assessment.

Custom Code Migration app on BTP — scans all Z-programs. Categories: clean, needs adjustment, needs refactoring, ABAP syntax error. At healthcare enterprise verbal estimate was 200 programs — tool found 640.

DBACOCKPIT column store analysis — actual HANA data volume in GB. Not estimates. Caught a 40% sizing error — proposed r6i.32xlarge, actual needed x2idn.32xlarge.

SAP Quick Sizer — SAPS rating from application workload. Validates AWS instance type selection.

Also run TAANA — Data Volume Management — identifies archiving candidates. Smaller migration volume means shorter SUM DMO runtime.

Then assess Unicode status, source ECC release level — minimum EHP7 for standard DMO, source database type, downtime tolerance, and custom code volume for migration approach decision.

Follow-up defense:

  • What happens if the source is below EHP7?
  • How do you handle non-Unicode source systems?

Q9. What migration approach do you choose — brownfield, greenfield, or SDT?

Brownfield — keep all config, all data, convert in place using SUM DMO. Right when custom code is manageable, business processes are mature, fastest path to S/4HANA.

Greenfield — fresh S/4HANA, redesign processes using SAP Best Practices. Right when ECC is heavily customized beyond remediation cost or customer wants clean slate.

Selective Data Transition — SDT — extract selected objects from ECC into fresh S/4HANA. Right when customer needs org restructuring, currency change, or consolidation of multiple ECC systems.

Decision is driven by four tool outputs — Readiness Check, Custom Code Migration app, DBACOCKPIT, Quick Sizer. Never by preference alone.

I always present all three options with trade-offs — never a single recommendation.

Follow-up defense:

  • When would you choose SDT over brownfield?
  • How does SDT handle controlling area consolidation?

Q10. Walk me through the SUM DMO technical execution.

SUM — Software Update Manager — with Database Migration Option.

Single orchestrated process — R3trans export from source ECC database, Unicode conversion if needed, HANA import on AWS target, S/4HANA conversion post-import.

Always run SUM PREPARE phase first — EXTRACTIONONLY mode. No data movement. Surfaces every blocker without executing migration.

Minimum two dress rehearsals. First gives actual SUM runtime and reveals ABAP syntax blockers. Second validates all fixes complete and downtime within agreed window.

Dress rehearsal environment provisioned on AWS from HANA Backint backup restored to temporary EC2 — gives most accurate runtime simulation.

At healthcare enterprise first rehearsal showed 22-hour runtime versus 20-hour window. Negotiated 28-hour window. Second rehearsal showed 18 hours after optimization. Cutover ran 18.5 hours — clean.

Cutover sequence — CTS freeze 72 hours before, final Backint backup, SUM DMO start, phase gate validation, post-conversion SPDD then SPAU, Fiori refresh /UI2/FLPD_CONF, BTP integration smoke tests, SM66 and ST05 checks, SNOTE, signed business go/no-go.

Follow-up defense:

  • What is the SUM EXTRACTIONONLY mode specifically?
  • How do you set up the dress rehearsal environment?

Q11. What is DMO versus DMO4 — when do you choose which?

DMO — standard Database Migration Option.

Source ECC on Oracle, SQL Server, DB2 — migrates to HANA and converts to S/4HANA in a single downtime window.

Right for database below 2 to 3 TB and downtime tolerance above 18 hours.

DMO4 — SUM DMO with System Move.

Designed for different source and target hosts — on-premises ECC to AWS target.

Pre-migration phase streams bulk data to AWS shadow system while source ECC stays live.

Only final delta sync and S/4HANA conversion require downtime — 4 to 8 hours total.

Right for database above 4 TB or downtime tolerance below 8 hours.

Decision matrix — database volume and business downtime tolerance are the two numbers that drive the decision.

Below 2 TB, above 18 hours — standard DMO. Above 4 TB or below 8 hours — DMO4. Above 8 TB with below 4 hours — DMO4 plus Near Zero Downtime option combined.

Follow-up defense:

  • What are the specific prerequisites for DMO4?
  • How does the replication log work in DMO4?

Q12. What are the prerequisites for DMO4?

SUM version 2.0 SP15 or higher — check SAP Note 2013521.

Source ECC minimum EHP5 SP13 — recommended EHP7 or EHP8.

SAP kernel 7.49 or higher on source application server.

Source must be Unicode — DMO4 does not support inline Unicode conversion.

Source database must meet minimum version — Oracle 11.2.0.4, SQL Server 2012, DB2 10.5.

AWS HANA target must be SAP-certified instance — validated against SAP Certified Hardware Directory.

Network — minimum 1 Gbps dedicated to SUM replication via Direct Connect — not internet VPN.

MTU 9001 jumbo frames on ENI.

Source system must have 20% free disk space for replication log.

All SAP add-ons must be compatible with target S/4HANA release — validated via SAP Maintenance Planner.

Business Partner migration must be complete before DMO4 starts.

Follow-up defense:

  • What happens if the replication log fills on source?
  • Why is Direct Connect mandatory — not VPN?

Q13. What is Downtime Optimized DMO versus Downtime Minimized Conversion — how do they work?

These are two different downtime reduction techniques. Complementary — not alternatives.

Downtime Optimized DMO — moves data transfer outside the downtime window.

Pre-migration phase streams bulk data to AWS shadow HANA while source ECC stays live.

Delta sync and S/4HANA conversion happen during downtime — 4 to 8 hours total.

Eliminates export, transfer, import phases from downtime — 18 to 26 hours removed.

Downtime Minimized Conversion — parallelizes the S/4HANA conversion steps inside the downtime window.

DDIC pre-activation runs on shadow system before downtime — during pre-migration phase.

S/4HANA application layer pre-installed on shadow.

Repository objects pre-generated for delta-independent objects.

Table conversions parallelized using HANA in-memory parallel execution — BSEG to ACDOCA, material ledger, business partner run simultaneously.

Combined result — standard DMO 28 to 35 hours reduced to 3 to 4 hours total downtime.

Requires SUM 2.0 SP17 or higher for Downtime Minimized Conversion. Check SAP Note 2881323.

Follow-up defense:

  • Why does HANA enable table conversion parallelization when Oracle cannot?
  • What is the shadow system and how is it managed?

Q14. How do you reduce SUM DMO runtime — what are the key levers?

Seven levers in order of impact.

Data archiving via TAANA and SAP ILM — 30 to 40% runtime reduction. Archive BSEG, MSEG, CDPOS before DMO. Every GB removed reduces import phase proportionally.

NZD — Near Zero Downtime — reduces production downtime to 4 to 8 hours. Pre-migrates bulk data before downtime window.

io2 Block Express storage plus parallel R3trans — upgrade HANA log and data volumes to io2 Block Express, increase ICNV_NO_CPUS parallel import processes based on CPU headroom from dress rehearsal.

Custom code remediation completion — every unresolved ABAP syntax error is a mid-run pause. At healthcare enterprise 47 unresolved errors caused 17 hours of dead pause time in run one. All resolved before run two — zero pauses.

Parallel post-conversion workstreams — pre-resolve all SPDD and SPAU conflicts from run one. Execute pre-made decisions in 20 minutes instead of 3 hours.

Dress rehearsal measurement — never commit to a downtime window without measured rehearsal runtime.

HANA parameter tuning — disable delta merge during import, set log_mode to overwrite, increase datashipping_parallel_channels to 4.

Follow-up defense:

  • How do you calculate exactly how many hours archiving saves?
  • What is the ICNV_NO_CPUS calculation?

Q15. How do you optimize the second SUM DMO run based on the first run?

Analyze DURATION.XML from run one immediately after completion.

Extract phase name, duration, retry count, status for every phase.

Sort by duration descending — top 10 phases are optimization targets.

Key findings from DURATION.XML:

  • Retry time versus net processing time per phase — dead time is the target.
  • Phase-level I/O correlation — cross-reference phase timestamps with CloudWatch EBSWriteOps, CPUUtilization.
  • Gap times between phases — network drops, manual interventions, lock waits.

Actions between run one and run two:

  • Archive data volumes identified from top table analysis.
  • Increase ICNV_NO_CPUS — calculated from CPU headroom measurement.
  • Upgrade storage to io2 Block Express if EBS IOPS was at ceiling.
  • Disable HANA delta merge during import.
  • Set log_mode to overwrite.
  • Pre-resolve all SPDD and SPAU conflicts.
  • Fix all ABAP syntax errors from run one conversion phase log.
  • Automate dress rehearsal environment reset via Step Functions — 4 hours instead of 2 days.

At healthcare enterprise — run one 28 hours total, run two 18 hours total. 10-hour reduction. Production cutover ran 18.5 hours.

Follow-up defense:

  • What exactly is in the DURATION.XML file?
  • How do you calculate the ICNV_NO_CPUS increase precisely?

Q16. What are SUM benchmarking tools and what do you get from them?

SUM has built-in DMO Runtime Estimator — runs in EXTRACTIONONLY mode.

Analyzes source system characteristics — export volume, repository objects, table conversions, parallel processes, network bandwidth.

Produces runtime estimate broken down by phase.

Third-party tools:

Panaya CloudQuality — custom code impact analysis and remediation effort estimation. Complexity scoring per program with estimated hours.

Tricentis Tosca — risk-based regression test optimization. Reduces test scope from 1,400 cases to 340 high-risk cases — saves 265 hours of test execution.

SNP CrystalBridge — data object level migration analysis. Most sophisticated for SDT scenarios — currency changes, controlling area consolidation. Used at telecom enterprise for ECC data model analysis.

Basis Technologies ActiveControl — transport dependency analysis. Reduces DDIC activation phase duration by preventing sequence conflicts.

Key outcome from benchmarking — quantified hours saved per optimization lever. Not guessing — precise calculation. Data archiving removes 600 GB — at 1 hour per 100 GB import rate — 6 hours removed from import phase. Exact number, not estimate.

Follow-up defense:

  • How does SUM EXTRACTIONONLY differ from a full dress rehearsal?
  • What does SNP CrystalBridge do that standard SAP SDT tooling cannot?

Q17. What is the DURATION.XML file and how do you use it?

DURATION.XML is written by SUM to the SUMLOG directory after every run.

Structured XML recording every SUM phase with start timestamp, end timestamp, duration in seconds, status, and retry count per attempt.

Key value — separates retry time from net processing time.

Standard MAIN_SHADOW.LOG shows phase status — does not cleanly separate retry time from actual processing.

DURATION.XML shows — DDIC_ACTIVATION ran 4.5 hours total, net processing 2.1 hours, retry time 2.4 hours from 8 activation conflicts. Wrong diagnosis without this — add parallel processes. Right diagnosis — pre-resolve DDIC conflicts.

Also use for cross-referencing with CloudWatch:

  • Extract phase timestamps from DURATION.XML.
  • Pull CloudWatch EBSWriteOps, CPUUtilization for exact phase time windows.
  • Gives per-phase resource utilization profile — not just overall run average.

At healthcare enterprise — found storage was bottleneck for first 6.3 hours of import, then delta merge consumed 45% CPU. Two different fixes for same phase.

Follow-up defense:

  • What tool do you use to parse DURATION.XML?
  • How do you correlate DURATION.XML timestamps with CloudWatch precisely?

Q18. What is the difference between DMO and DOC?

DMO — Database Migration Option.

Migrates ECC to S/4HANA while simultaneously changing the database engine from Oracle, SQL Server, or DB2 to HANA.

Two things in one SUM run — database migration and application conversion.

Used when source is on a non-HANA database.

DOC — Database Migration Option for Conversion.

Used when source ECC is already on HANA — HANA TDI on-premises or EC2.

Database engine is already HANA — no database migration needed.

SUM performs only the S/4HANA application conversion — not a database migration.

Single-phase process — faster than DMO because no R3trans export and import cycle.

Decision — what database is the source ECC running on. Non-HANA source — DMO. HANA source — DOC or HANA System Replication path.

Follow-up defense:

  • If source is ECC on HANA on EC2, what is the fastest path to S/4HANA on RISE?

Q19. Have you used SWPM for classical SAP migration?

Yes. SWPM — Software Provisioning Manager — replaced the old sapinst tool from SAP Basis 7.40.

Used it at a healthcare enterprise for heterogeneous system copy.

Source — ECC on Oracle on Linux on-premises. Target — ECC on HANA on AWS — r6i.16xlarge.

SWPM handled the OS and database platform change in a single provisioned run.

Also used at global consulting firms for new system installations — ABAP stack, Java stack, dual stack — and system copies for sandbox and development environments.

Classical migration using SWPM uses R3load as the underlying data movement engine — different from SUM DMO which uses R3trans.

R3load runs parallel export and import streams simultaneously — significantly faster for bulk data migration.

SWPM orchestrates R3load export on source, transfer to target, R3load import into HANA, post-installation configuration.

Key difference from SUM DMO — SWPM copies the system at database level first. Upgrade to S/4HANA is a separate subsequent step. Two events — copy then upgrade. SUM DMO does both in one run.

Follow-up defense:

  • When would you use SWPM over SUM DMO?
  • How does R3load differ from R3trans technically?

Q20. Walk me through RISE with SAP migration on AWS technically.

RISE means SAP operates S/4HANA Cloud Private Edition in SAP-managed AWS accounts. SAP manages infrastructure, OS, HANA, and SUM execution. I own preparation, connectivity, integration, data migration, and validation.

Connectivity — Transit Gateway or VPC Peering from customer AWS VPC to SAP-managed RISE VPC. RFC ports 33{NN} and HTTPS 443. Cloud Connector reverse tunnel for on-premises systems.

Identity — SAP IAS federating with customer Entra ID or Okta via SAML 2.0. IAS as proxy IdP for S/4HANA, BTP, SAP Build Work Zone.

Data migration — LTMC for standard objects, LTMOM for custom objects. For large datasets — extract to S3, AWS Glue ETL to LTMC format, upload via RISE SFTP or Cloud Data Services.

Pre-upgrade — SAP Maintenance Planner stack XML, Custom Code Migration app BTP analysis, CTS freeze 72 hours, SUM simulation in EXTRACTIONONLY via RISE Support Portal ticket.

Post-upgrade — SPDD then SPAU, Fiori refresh /UI2/FLPD_CONF, BTP iFlow smoke tests, OData version validation, SM66 and ST05 baseline, SNOTE before business user access.

Follow-up defense:

  • What ports are required for RISE connectivity?
  • How does RISE upgrade patching work — who owns what?

Q21. If you have to migrate ECC on EC2 to S/4HANA on RISE — what is your approach?

First question — what database is the source ECC running on EC2.

Scenario A — Non-HANA database on EC2 — Oracle, SQL Server, DB2.

Approach — SUM DMO from EC2 to RISE.

EC2 source already on AWS — data transfer over AWS internal network instead of Direct Connect.

AWS internal at 25 Gbps — 2 TB database transfers in 12 to 15 minutes versus 1.5 to 2 hours over Direct Connect.

Configure Transit Gateway between customer AWS account and SAP-managed RISE AWS account.

Dress rehearsal environment — EC2 AMI snapshot, launch from AMI — ready in under 1 hour versus 4 hours for on-premises.

Scenario B — HANA database on EC2 — HANA TDI on EC2.

Fastest possible path — same database engine both sides.

Use HANA System Replication to replicate source HANA to RISE HANA target.

HSR runs in background while ECC stays live — initial sync in 2 to 4 hours over AWS backbone.

Downtime is only HSR takeover plus S/4HANA application conversion — 3.5 to 5.5 hours total.

No R3trans export, transfer, or import — dramatically shorter downtime than standard DMO.

Follow-up defense:

  • What Transit Gateway configuration is needed for RISE connectivity?
  • What if the source HANA version is below the HSR compatibility minimum?

Q22. What is Clean Core and ABAP RAP in S/4HANA?

Clean Core is SAP's principle — no customer logic inside the SAP core software components.

All custom logic must use SAP-published extension points — in-app via released APIs or side-by-side on BTP.

Reason — upgrade safety. Every core modification risks conflict with next SAP support package.

Three extension types:

In-app extensions using ABAP RAP — ABAP RESTful Application Programming model.

Runs inside S/4HANA ABAP system. Uses only released SAP APIs, CDS view entities, and released ABAP language constructs. No direct database table access. No deprecated function modules. No internal SAP APIs.

Release state configured in SE11 or ABAP Development Tools in Eclipse.

Side-by-side extensions on BTP using CAP — Cloud Application Programming model.

Logic runs on BTP. Communicates with S/4HANA via published OData V4 APIs. BTP Connectivity service reaches S/4HANA through Cloud Connector.

Key user extensions — zero-code additions using SAP Fiori Custom Fields and Logic app. No ABAP development. Configure extension fields on standard business objects — Sales Order, Purchase Order, Business Partner.

Follow-up defense:

  • How do you assess Clean Core compliance for an existing system?
  • What is the difference between RAP and a classic BADI?

Q23. What is HANA System Replication and how do you set it up between primary and secondary?

HSR replicates HANA persistence — data volumes and log volumes — from primary to secondary continuously.

Synchronous mode — primary waits for secondary acknowledgment before committing transaction. RPO zero.

Setup steps:

Copy SSFS PKI files from primary to secondary — identical files required for trusted channel. md5sum SSFS_SID.DAT on both nodes must match. Most common HSR setup failure when skipped.

Stop HANA on secondary:

HDB stop

Register secondary to primary:

hdbnsutil -sr_register \
  --remoteHost=primary_hostname \
  --remoteInstance=00 \
  --replicationMode=sync \
  --operationMode=logreplay \
  --name=SECONDARY_AZ2 \
  --remoteName=PRIMARY_AZ1

Start HANA on secondary:

HDB start

Initial full data sync starts automatically — primary ships entire data volume to secondary.

Validate:

hdbsql "select SITE_NAME, REPLICATION_STATUS, SHIPPED_LOG_POSITION, REPLAYED_LOG_POSITION from M_SYSTEM_REPLICATION"

REPLICATION_STATUS must show synced.

Follow-up defense:

  • What is the difference between sync, syncmem, and async replication modes?
  • What is logreplay operationMode and why is it preferred?

Q24. What are all the components needed for SAP application-level HA?

ASCS — ABAP Central Services — message server and enqueue server. Single point of failure — requires Pacemaker cluster failover to standby instance.

ERS — Enqueue Replication Server — continuously replicates enqueue lock table from ASCS. Always on opposite instance from ASCS. On failover they swap — ASCS moves to instance 2, ERS moves to instance 1.

PAS — Primary Application Server — stateless, connects to ASCS and HANA. HA via multiple application servers.

AAS — Additional Application Servers — spread across AZs. EC2 Auto Scaling Group manages lifecycle. Message server load balancing removes failed AAS from pool automatically.

SAP Web Dispatcher — ALB in front, Web Dispatcher in Auto Scaling Group minimum 2 for HA within region.

/sapmnt EFS — shared file system. AWS EFS with multi-AZ mount targets — single EFS accessible from all AZs in production region.

Overlay IP — floating virtual IP for HANA always pointing to current primary. Application servers connect via overlay IP only.

Pacemaker cluster — manages ASCS failover, ERS swap, overlay IP movement, stonith fencing.

Route 53 health checks — external monitoring of HANA and ASCS virtual hostnames.

Follow-up defense:

  • Why must ASCS and ERS always be on different instances?
  • What happens to active locks if ASCS fails without ERS?

Q25. What are the database components for always-on availability?

Eight layers required for always-on HANA availability.

Layer 1 — HANA System Replication — data replication foundation. Sync mode, zero RPO.

Layer 2 — Pacemaker Cluster Resource Manager — SAPHanaTopology, SAPHana, IPaddr2, and Stonith resource agents. Monitors health and executes automatic failover.

Layer 3 — Stonith Fencing — fence_aws agent. Prevents split-brain by forcibly stopping failed node before secondary promotes. Never skip stonith testing.

Layer 4 — Overlay IP — virtual IP from VPC CIDR. Always assigned to current primary. Application servers connect to overlay IP only. Moves via route table update or Route 53 on failover.

Layer 5 — Route 53 health checks — TCP on HANA port, 10-second interval, 3-failure threshold. Alerts on HANA availability issues invisible to Pacemaker.

Layer 6 — AWS Backint Agent backup — full weekly plus 15-minute log backups to S3. Protects against logical corruption that HSR replicates to both nodes.

Layer 7 — HANA Watchdog — internal HANA process health monitor. Triggers controlled shutdown on repeated nameserver crashes — faster Pacemaker detection than undetected crash.

Layer 8 — CloudWatch monitoring — HANA memory, dialog response time, work process depth, HSR replication status, HSR replication lag. Alert if HSR not synced — RPO zero window is open.

Follow-up defense:

  • What happens in split-brain without stonith configured?
  • Why is the overlay IP not a secondary private IP in cross-AZ HA?

Q26. Can you choose an overlay IP from the private subnet where HANA is deployed?

Yes — but with an AWS-specific constraint that changes the design from on-premises.

On-premises — overlay IP is any unused IP in the network segment. Same broadcast domain — all nodes see the IP immediately.

On AWS — primary and secondary HANA instances are in different AZs — always different subnets.

Primary in 10.0.1.0/24 — secondary in 10.0.2.0/24.

AWS enforces that secondary private IPs must come from the subnet the ENI belongs to.

Cannot assign an IP from 10.0.1.0/24 to an instance in 10.0.2.0/24.

Solution — choose overlay IP from outside both instance subnets.

Use a dedicated /28 CIDR block — example 10.0.3.0/28 — reserved for overlay IPs.

Or an IP from VPC CIDR not assigned to any subnet.

Configure a VPC route table entry — destination overlay_ip/32, target current primary ENI ID.

On failover — Pacemaker executes aws ec2 replace-route — changes target to new primary ENI.

Route propagates across VPC in under 1 second.

At healthcare enterprise I used route table manipulation — 1-second failover versus 30 to 60 seconds for Route 53 DNS approach.

Follow-up defense:

  • What IAM permissions does the EC2 instance profile need for route table update?
  • Can the overlay IP be in a different VPC CIDR block entirely?

Q27. Walk me through DR setup for SAP on AWS — database and application server sides.

First I ask four questions — RTO, RPO, DR tier, budget.

Database DR — HANA async replication to DR region.

Multi-tier HSR topology:

  • Node 1 — Production Primary — AZ1 — us-east-1a — sync HA to Node 2.
  • Node 2 — Production Secondary — AZ2 — us-east-1b — async DR to Node 3.
  • Node 3 — DR HANA — us-west-2a.

Replicate from Node 2 not Node 1 — offloads DR replication overhead from primary. Primary performance not impacted.

Cross-region connectivity — Transit Gateway inter-region peering or VPC Peering. AWS backbone — no internet exposure.

S3 Cross-Region Replication — production backup bucket to DR backup bucket. S3 Replication Time Control — 99.99% of objects replicated within 15 minutes.

Application Server DR — Pilot Light model.

HANA DR running continuously. Application servers stopped — represented by EC2 AMIs and CloudFormation templates.

Weekly automated AMI refresh via SSM Automation — copies AMIs to DR region. AMI IDs stored in Parameter Store. CloudFormation references Parameter Store.

EFS Replication — production EFS to DR EFS — continuous, sub-1-minute RPO. DR application servers mount DR EFS for /sapmnt.

Route 53 failover routing — health check on production ALB, automatic DNS switch to DR ALB on failure.

On DR declaration — launch CloudFormation stacks in sequence: ASCS, ERS, PAS, AAS. Total application server activation — 10 to 15 minutes. Combined with HANA DR takeover — total RTO 15 to 30 minutes for warm standby.

Follow-up defense:

  • How do you validate DR quarterly without impacting production?
  • What is the RPO difference between HSR async and S3 CRR backup?

Q28. If application servers are pre-deployed in DR — how do you ensure /sapmnt is synchronized?

Primary mechanism — AWS EFS Replication from production EFS to DR EFS.

Continuous replication — sub-1-minute RPO for SAP profile changes and kernel updates.

DR EFS is read-only while replication is active.

Pre-deployed DR application servers mount DR EFS at /sapmnt — see current production profiles.

CloudWatch monitoring:

  • MetricName — TimeSinceLastSuccessfulReplication.
  • Alert if lag exceeds 300 seconds — 5 minutes.
  • SNS notification to operations team.

Additional validation — Lambda sentinel file check every 15 minutes. Writes current timestamp to /sapmnt/SID/global/DR_SYNC_CHECK.txt on production EFS. Reads same file on DR EFS — validates timestamp within expected lag. If mismatch exceeds threshold — CloudWatch custom metric EFS_SYNC_STATUS = 0 — alarm fires.

On DR declaration:

  • Break EFS replication — aws efs delete-replication-configuration. DR EFS becomes read-write in 30 to 60 seconds.
  • Remount DR EFS as read-write on all DR application servers.
  • Run profile update script — replace production HANA hostname with DR HANA overlay IP in DEFAULT.PFL.
  • Start SAP processes in sequence — ASCS, ERS, PAS, AAS.

Follow-up defense:

  • What is the initial sync approach for large /sapmnt filesystems?
  • How do you handle SAP kernel patch synchronization?

Q29. Have you set up HANA database backup?

Yes — at healthcare enterprise I owned the complete HANA backup architecture on AWS.

AWS Backint Agent for SAP HANA — AWS-provided binary implementing SAP Backint interface specification.

Streams backup directly from HANA to S3 using multipart upload. Never file-based backup for production on AWS.

Installation — /usr/sap/SID/SYS/global/hdb/opt/aws-backint-agent/

Configuration file — aws-backint-agent-configuration.yaml — region, bucket, prefix, kms_key_id, compress=true.

HANA global.ini configuration:

catalog_backup_using_backint = true

data_backup_parameter_file and log_backup_parameter_file pointing to Backint Agent config.

Without catalog_backup_using_backint=true — backup catalog stays on local disk. Instance failure means catalog loss — cannot identify S3 objects for restore.

IAM instance profile permissions — s3:PutObject, s3:GetObject, s3:ListBucket scoped to specific backup bucket ARN and prefix. kms:GenerateDataKey, kms:Decrypt scoped to specific KMS key ARN. No wildcards.

Schedule — full weekly Sunday 01:00 AM, delta daily, log backup every 15 minutes.

log_mode = normal — enables continuous log backup for point-in-time recovery.

Monthly restore test — restore to temporary EC2, validate HANA starts cleanly, validate record counts. Untested backup is not a backup.

Follow-up defense:

  • What happens if log_mode is set to overwrite?
  • Why must catalog backup also go to S3 via Backint?

Q30. Have you done HANA HA and DR setup for a customer?

Yes — at healthcare enterprise I designed and implemented complete HA and DR.

HA — HANA System Replication synchronous mode across two AZs. Pacemaker with SAPHanaTopology, SAPHana, IPaddr2, and fence_aws stonith resources. Overlay IP via VPC route table manipulation. HSR replication lag consistently under 1 second on production workload. Validated automatic failover in controlled window — 47 seconds end-to-end from primary failure to application server reconnection. Zero data loss confirmed.

DR — Multi-tier HSR. Production primary to production secondary sync HA. Production secondary to DR region async — cross-region via VPC Peering. S3 CRR for backup replication — confirmed 8-minute average replication to DR bucket. Pilot Light model — only HANA EC2 running in DR region. CloudFormation templates for application server launch on DR declaration. Quarterly DR drills — consistent 18-minute RTO achieved versus 6-hour target.

HIPAA controls — EBS KMS-CMK, CloudTrail log file validation, S3 Object Lock Compliance 6-year retention per 45 CFR 164.316. Zero HIPAA findings post go-live.

Follow-up defense:

  • Walk me through your stonith fence_aws configuration.
  • What was your HANA global.ini system_replication parameter tuning?

Q31. What is HANA tenant copy and how is it used?

SAP HANA from version 2.0 uses Multi-Tenant Database Container — MDC architecture.

One HANA instance — System Database — hosts multiple independent tenant databases.

Each tenant has separate schema, users, memory, backup catalog, configuration.

Tenant copy is copying one HANA tenant database — within same HANA system or to a different HANA system — as a complete self-contained unit.

Use cases — system refresh from production to QA, migration to AWS, landscape consolidation, DR testing, performance testing, data masking for development.

Three technical methods:

Method 1 — Tenant backup and restore. Full backup of source tenant via Backint to S3. Restore as new tenant on target HANA system — HANA RECOVER DATABASE command. Most common. Works across HANA systems and regions.

Method 2 — HANA tenant move. hdbnsutil -sr_state — validate source is healthy. ALTER SYSTEM MOVE TENANT command — moves tenant from source to target HANA on same or different system. Requires HANA system replication or shared storage between source and target. Faster than backup-restore for large tenants.

Method 3 — HANA System Replication — used for migration not just copy. Replicate entire source HANA to target, then takeover. Used in ECC on HANA to S/4HANA on RISE scenario.

Follow-up defense:

  • How do you rename a tenant after copy?
  • What are the MDC restrictions for tenant copy across different HANA revisions?

Q32. What is your approach when a customer asks to migrate SAP ECC to S/4HANA with tight timeline?

First — I do not take a verbal timeline at face value.

Four tool outputs determine the real timeline — SAP Readiness Check 2.0, Custom Code Migration app on BTP, DBACOCKPIT column store analysis, Quick Sizer.

Without running these tools first — any timeline commitment is a guess.

At healthcare enterprise verbal estimate was 6 months. Custom Code Migration app found 640 programs versus 200 estimated — extended by 8 weeks. That extension was data-driven, not a surprise.

For timeline reduction — seven levers. Reduce data volume first — TAANA archiving before migration. Use Downtime Optimized DMO plus Downtime Minimized Conversion — shrinks production downtime to 3 to 4 hours. Pre-resolve SPDD and SPAU from dress rehearsals — eliminates dead time from post-conversion activities. Complete custom code remediation before SUM starts — zero mid-run pauses. Parallel workstream execution — infrastructure build, integration redesign, master data prep run simultaneously. Automate dress rehearsal environment setup — Step Functions reduces 2-day setup to 4 hours. Use Tricentis Tosca for regression test optimization — reduces 1,400 test cases to 340 high-risk cases.

Follow-up defense:

  • What is the minimum realistic timeline for a 2 TB brownfield S/4HANA migration?
  • How do you communicate timeline changes to stakeholders?

Q33. Walk me through RISE with SAP pre and post upgrade activities.

Pre-upgrade — I own all preparation. SAP owns SUM execution.

SAP Maintenance Planner — generate stack XML. Validate against product compatibility matrix — S/4HANA release, BW Add-On, Fiori Frontend Server, SAP_BASIS kernel.

Custom Code Migration app on BTP — simplification item delta analysis for the specific upgrade release pair. Resolve all red items before upgrade window. At healthcare enterprise pre-check caught 47 programs — all resolved before SUM started.

HANA revision certification check — HANA Hardware Directory cross-reference for specific AWS instance type. Not every HANA revision certified on every instance. Check HANA Platform Availability Matrix.

CTS transport freeze 72 hours before. Full HANA Backint backup to S3. EC2 AMI snapshot — dual rollback path.

SUM simulation in EXTRACTIONONLY — request via RISE Support Portal ticket. SAP runs simulation, provides runtime estimate and blocker list.

Post-upgrade — I own validation.

SPDD then SPAU — dictionary first, workbench second. Both clean before functional testing.

Fiori /UI2/FLPD_CONF catalog refresh. SAP Fiori Apps Reference Library for new release.

BTP iFlow monitor — check failed messages during upgrade window. OData version validation — V2 to V4 deprecation check.

SM66 and ST05 — performance baseline re-establishment.

SNOTE — master SAP Note plus Rapid Response Notes for target release before business user access.

Follow-up defense:

  • What is the RISE patching cadence?
  • Why is SPDD run before SPAU?

Q34. What SAP BTP integration patterns connect SAP and AWS?

Pattern 1 — OData V4 API to AWS Lambda via API Gateway.

S/4HANA OData V4 event → BTP iFlow with message mapping and dead-letter S3 → AWS API Gateway → Lambda → DynamoDB / SNS / SQS.

Pattern 2 — SAP Event Mesh to Amazon EventBridge.

S/4HANA SWFVISU event config → SAP Event Mesh AMQP → BTP iFlow bridge → EventBridge custom bus → Lambda / Step Functions / SQS.

Full decoupling — S/4HANA unaware of AWS targets.

Pattern 3 — BTP Data Intelligence to S3 and Glue.

S/4HANA CDS views via OData → BTP Data Intelligence pipeline → S3 Parquet → Glue Catalog → Athena / Redshift Spectrum.

SAP operational data flows to AWS data lake without exposing HANA ports.

Pattern 4 — Bedrock RAG for SAP invoice exception intelligent processing.

SAP workflow RFC → BTP iFlow → S3 PDF → Lambda → Textract with TABLES and FORMS → Bedrock InvokeModel with Claude plus FAISS/EFS RAG → JSON → BTP iFlow → SAP Invoice Management OData API. Human review if confidence below 0.75.

Follow-up defense:

  • How do you handle BTP iFlow dead-letter processing?
  • What is the SWFVISU transaction and what does it configure?

Q35. How do you design automation in an SAP landscape?

Layer 1 — SAP Build Process Automation on BTP.

Approval workflows and exception routing. Trigger as REST API from S/4HANA via ABAP HTTP client class CL_HTTP_CLIENT. Monitor in BTP Process Visibility workbench — cycle time metrics per step.

Layer 2 — UiPath RPA with SAP GUI Scripting API.

For legacy systems without API surfaces. Enable SAP GUI scripting — sapgui/user_scripting=TRUE in instance profile. Exception handler writes structured failure record to SQL Server — bot_name, transaction, error_code, retry_count. Power BI real-time monitoring dashboard on exception log table.

At healthcare enterprise — 24 bots deployed on $300K budget, $500K annual savings. Most complex bot — 6-step AP reconciliation, 3.5 hours daily manual reduced to 18 minutes automated.

Layer 3 — Textract plus Bedrock for intelligent document automation. Described in invoice exception use case above.

Layer 4 — Migration tooling automation.

CloudFormation for SAP landscape provisioning. Step Functions for automated dress rehearsal restore from Backint S3 — reduced setup from 2 days to 4 hours. PyRFC-based post-cutover validation script — predefined business process checks, pass/fail per check.

Follow-up defense:

  • How do you handle bot exception escalation in UiPath?
  • What is the PyRFC library and how does it connect to SAP?

Q36. How do you manage security and HIPAA compliance for SAP on AWS?

Network layer — HANA ports 3{NN}13 and 3{NN}15 restricted to application tier subnet CIDR only via security groups. AWS Config custom rule — alerts within 15 minutes if any HANA port opens to broader CIDR.

Identity layer — IAM instance profiles, no stored access keys. HANA backup instance profile — s3:PutObject scoped to specific bucket ARN and prefix only.

HIPAA specific:

  • EBS volumes encrypted with KMS customer-managed keys — CMK rotated annually.
  • CloudTrail with log file validation enabled — detects any audit log tampering.
  • VPC flow logs shipped to dedicated security account via cross-account S3 replication.
  • S3 Object Lock Compliance mode — SAP Security Audit Log bucket — 6-year retention per 45 CFR Part 164.316.
  • AWS Macie on S3 buckets for SAP data exports — detects unintended PHI exposure.

SAP application controls:

  • rsau/enable=1 and rsau/selection_slots=10 in instance profile — activates SAP Security Audit Log.
  • SM20 audit log exported to CloudWatch Logs — 6-year retention.
  • Daily SU53 authorization failure digest alert via ABAP batch job.
  • CTS-only production access — no developer has direct production access.
  • STMS import log integrated with AWS CodePipeline — every production transport auditable.

Follow-up defense:

  • What specific PHI tables did you identify in S/4HANA for GDPR erasure?
  • How did you implement the right-to-erasure for customer master data?

Q37. How do you handle cost optimization for SAP on AWS?

Production SAP — 1-year Reserved Instances or Savings Plans for HANA production instances. 30 to 40% discount on on-demand.

Non-production SAP — EC2 Instance Scheduler to automatically stop all non-production systems outside business hours. 12-hour shutdown daily plus full weekend shutdown. For 3 non-production SAP systems — 60 to 70% non-production compute cost reduction.

Smaller instance types for non-production HANA — r6i.8xlarge or r6i.16xlarge — not production size.

S3 Intelligent-Tiering on Backint backup bucket — automatically moves backups older than 90 days to infrequent access. Approximately 40% backup storage cost reduction.

Right-sizing — CloudWatch HANA memory utilization over previous 90 days. Average below 50% — instance is likely oversized.

Quarterly cost optimization review — right-size check, scheduling check, data lifecycle check. Not one-time — SAP landscapes change and what was right-sized at go-live is frequently wrong 12 months later.

Built 5-year TCO model at healthcare enterprise — cloud 23% cheaper year one, 70% cheaper year five versus on-premises hardware refresh. Cloud decision made on data — not preference.

Follow-up defense:

  • How do you handle reserved instance commitments if the project scope changes?
  • What CloudWatch metrics drive your right-sizing recommendation?

Q38. What is Selective Data Transition and when do you recommend it?

SDT is the third migration approach alongside brownfield and greenfield.

Lets customers migrate selected objects from ECC to S/4HANA while leaving others behind or restructuring organizational model.

Right choice when customer needs to: consolidate multiple ECC systems into single S/4HANA instance, split single ECC into multiple S/4HANA instances, change currency or fiscal year variant — locked in brownfield conversion, migrate only specific company codes while leaving others on ECC temporarily.

Technical execution:

  • Install fresh greenfield S/4HANA on AWS target.
  • Use SDT extraction tools to extract selected objects from source ECC in SDT XML format — object-level selectivity by company code, plant, profit center.
  • Load extracted objects into S/4HANA through SDT import tool.
  • Financial data uses SDT financial migration object for G/L balance migration with correct posting period alignment.

Highest complexity scenario — currency change plus controlling area consolidation.

Mapping tables for cost center hierarchy, profit center assignment, internal order structure must be configured and validated in dry run before production extraction.

Profit center mapping error causes every migrated FI document to post to wrong segment — manual journal entry correction required.

Follow-up defense:

  • How does SDT differ technically from a greenfield with data migration?
  • What is the SDT extraction format and what tools generate it?

Q39. How do you approach customer requirements gathering and infrastructure assessment for SAP migration?

Three layers — business layer, landscape layer, technical depth layer.

Business layer — start with what the customer is trying to achieve — not what they want to migrate. What business outcomes drive this? Faster financial close, real-time inventory, procurement cost reduction? What has already been tried and failed?

Landscape layer — SAP system landscape overview. ECC version and SP level, active modules, user count and geography, custom code Z-program count, external integrations list. One structured questionnaire plus 90-minute technical discovery call with customer Basis team.

Technical depth layer — four tool outputs — SAP Readiness Check 2.0, TAANA data volume, Custom Code Migration app BTP, SAP Quick Sizer. These give the real migration complexity picture — not what the customer thinks they have.

Based on three layers — recommend migration approach with risk-adjusted timeline. Always two to three options with trade-offs — customer makes an informed decision, not a salesperson decision.

Follow-up defense:

  • What is in your standard discovery questionnaire?
  • How do you handle customers who refuse to run the discovery tools?

Q40. How do you manage risk throughout a large SAP migration program?

Three levels — pre-migration identification, in-flight monitoring, cutover mitigation.

Pre-migration — risk register with three categories. Technical risk — custom code volume, integration complexity, HANA sizing certainty. Delivery risk — resource availability, customer team involvement, vendor contracts. Business risk — process change readiness, parallel go-live with business events, regulatory deadlines. Each risk gets probability, impact score, owner, mitigation action. Reviewed every steering committee.

In-flight — three leading indicators tracked every sprint:

  • Custom code remediation velocity — percentage of Z-programs fully remediated versus plan.
  • Integration test pass rate — BTP iFlows passing end-to-end tests.
  • Data migration success rate — records loaded cleanly in most recent dry run.

Two consecutive sprints below plan on any indicator — flag as program-level risk, propose corrective action.

Cutover — pre-decided decision trees. SUM DMO hits 80% of window at 60% completion — rollback. No real-time debate. Post-conversion finds critical FICO posting defect — rollback. No go-live exception. Document these trees and get business and IT leadership to sign before cutover window opens.

Follow-up defense:

  • How do you handle a stakeholder who pushes back on your risk assessment?
  • What happened when you had to invoke a rollback decision?

Q41. What is your weakness or area for growth?

Earlier I moved fast on technical decisions without leaving enough documentation trail.

Decision was right but three months later team members could not explain a HANA instance type selection to the customer CFO — because only I had the full reasoning.

Feedback was direct — team trusted my judgment but could not represent decisions when I was not in the room.

I changed my process. I now write a one-page Architecture Decision Record for every significant decision — the decision, alternatives considered, trade-offs, risks accepted, conditions for revisiting. Done within 24 hours. Takes 20 to 30 minutes. Team can represent any architecture choice to any stakeholder at any time.

I still move fast. Documentation discipline is now part of how I move fast.

Q42. What questions do you have for the interviewer?

Three senior-level questions.

What does success look like for a new Delivery Consultant in the first six months — is it landing the first customer engagement, building reusable assets, or establishing pre-sales credibility with account teams?

How does the organization balance depth — closely embedded in one SAP engagement end to end — versus running multiple engagements in parallel? What does the typical engagement load look like at senior level?

Where are the two or three most technically unsolved challenges in SAP migration work right now — areas where a senior practitioner joining today could make the most immediate contribution?

Related Interview Support

Need live support during your SAP interview? ProxyTechSupport provides real-time answer generation during live technical interviews.

WhatsApp: +91 96606 14469
https://wa.me/919660614469

Services:

  • Real-time SAP technical interview support
  • Live answer delivery via screen share
  • SAP migration proxy interview assistance
  • AWS architecture interview support
  • BTP integration interview preparation
  • HANA HA/DR interview coaching
  • Resume and JD alignment preparation

Related interview guides:

For preparation, see our proxy interview support page and our AWS job support page for technical pre-interview coaching.

FAQ Section

What is SUM DMO in SAP migration?

SUM — Software Update Manager with Database Migration Option — is the SAP-provided tool that migrates ECC from Oracle, SQL Server, or DB2 to HANA while simultaneously converting the application to S/4HANA in a single downtime window.

What is the difference between brownfield and greenfield S/4HANA migration?

Brownfield converts existing ECC in place using SUM DMO — preserving all configuration, data, and custom code. Greenfield starts fresh on S/4HANA using SAP Best Practices. Brownfield is faster with less disruption. Greenfield allows process redesign from scratch.

What is RISE with SAP on AWS?

RISE with SAP is SAP's subscription model where SAP operates S/4HANA Cloud Private Edition in SAP-managed AWS accounts. SAP manages infrastructure, OS, HANA, and upgrade execution. Customer owns preparation, connectivity, integration, data migration, and validation.

How does HANA System Replication work for HA?

HSR replicates HANA persistence — data volumes and log volumes — from primary to secondary node continuously. Synchronous mode means primary waits for secondary acknowledgment before committing transactions — zero data loss. Pacemaker cluster resource manager executes automatic failover when primary fails — typically 47 to 90 seconds end to end.

What is Downtime Optimized DMO?

Downtime Optimized DMO pre-migrates bulk data to the AWS target while source ECC stays live. Production downtime is reduced to only the final delta sync plus S/4HANA conversion — 4 to 8 hours instead of 20 to 35 hours for standard DMO.

What is Clean Core in S/4HANA?

Clean Core means no customer custom logic lives inside the SAP core software components. All custom logic must use SAP-published extension points — in-app via ABAP RAP with released APIs or side-by-side extensions on BTP using CAP. This ensures upgrade safety — no conflicts with SAP support packages.

What is the DURATION.XML file in SUM?

DURATION.XML is generated by SUM in the SUMLOG directory after every run. It records every phase with start timestamp, end timestamp, duration in seconds, status, and retry count. Used to identify which phases consumed the most time and how much was dead time from retry cycles versus actual processing — drives optimization between dress rehearsal runs.

What is SAP BTP Integration Suite?

BTP — Business Technology Platform — Integration Suite is SAP's cloud-native integration middleware replacing SAP PI/PO. It uses iFlow-based orchestration to connect S/4HANA to AWS services — API Gateway, EventBridge, Lambda, Bedrock — and to other SAP and non-SAP systems.

How do you set up DR for SAP on AWS?

DR uses multi-tier HANA System Replication — sync HA replication within the production region, async DR replication to a second AWS region. Application servers use Pilot Light model — HANA EC2 running in DR region, application servers launched from pre-built AMIs via CloudFormation on DR declaration. EFS Replication keeps /sapmnt synchronized between regions continuously.

What are the prerequisites for DMO4 — SUM with System Move?

SUM 2.0 SP15 or higher, source ECC minimum EHP5, SAP kernel 7.49+, source must be Unicode, source database minimum version, target must be SAP-certified AWS instance, minimum 1 Gbps dedicated Direct Connect bandwidth, MTU 9001 jumbo frames on ENI, 20% free disk on source for replication log, all add-ons compatible with target S/4HANA release, Business Partner migration complete.

Bottom CTA

Need real-time SAP interview support?

WhatsApp ProxyTechSupport now:

+91 96606 14469
https://wa.me/919660614469

We support SAP migration interviews, AWS architecture interviews, SAP Basis/HANA interviews, BTP integration interviews, cloud migration job support, live technical interview support, coding test support, and production issue support. See our proxy interview support page for full service details.