Infor M35 min readNetray Engineering Team

Infor M3 Upgrade and CloudSuite Migration Planning

Migrating Infor M3 to CloudSuite is less a technical upgrade than a controlled removal of everything that made your on-premises instance unique. Multi-tenant CloudSuite does not accept business engine modifications, direct database access, or custom jobs on the application server, so every one of those must be redesigned as configuration, personalization, API integration, or an external service. Most mid-size manufacturers complete the journey in nine to eighteen months. The variable is not technology, it is how many modifications and integrations you accumulated and how honestly you inventory them at the start.

Inventory Modifications, Interfaces, and Direct Database Access

Start with a complete register: every business engine modification, every custom program, every interface in or out, every report built on direct SQL, every scheduled script on the application server, and every H5 or Smart Office script. For each entry capture the business purpose, the owner, the last time it actually ran, and whether an equivalent capability now exists in standard M3. Landscapes running ten or more years commonly find that a meaningful share of modifications are dead, superseded by standard functionality delivered in later releases, or maintained for a business process the company abandoned. Retiring those is the cheapest scope reduction available.

  • Instrument the current system to prove which modifications and reports are actually used before deciding
  • Classify each item as retire, replace with standard, reconfigure, or rebuild as an extension
  • Assume every direct database write must be rebuilt against MI API transactions
  • Convert on-server scheduled scripts to externally hosted services calling APIs

Rebuilding Customizations as Cloud Extensibility

The cloud extensibility toolkit is genuinely capable once teams stop trying to reproduce a modification line for line. Field-level defaults, hidden fields, and guided flows become H5 personalization and scripting. Cross-application document flows become ION with business object documents. Event-driven logic becomes Event Hub subscribers running in an external service. Document output and archiving moves to Infor Document Management. Reporting moves off direct SQL to the cloud reporting and data lake tooling. The rule of thumb worth adopting early: if a requirement cannot be met by configuration, personalization, an API, or an event subscriber, challenge the requirement before designing around it.

  • Rewrite requirements as outcomes, not as descriptions of the old modification
  • Move batch logic into external services that call MI APIs on a schedule you control
  • Replace direct SQL reports with data lake or reporting tooling before the migration, not during
  • Budget rework effort per modification rather than a flat percentage; complexity varies enormously

Data Migration and History Strategy

Deciding how much history to carry is the second largest cost driver. Full transactional history migration is slow, expensive to validate, and rarely justified. The common pattern is to migrate open items and master data through APIs, load a defined window of closed transactions for comparability, and keep older history in a read-only archive or data lake that finance and audit can query. Plan at least three full mock conversions: the first to prove the mechanism, the second to measure runtime and fix quality, and the third as a rehearsal on production-scale data with the actual cutover team, timings, and rollback decision points documented.

Testing Waves, Cutover, and Life After Go-Live

Structure testing in waves: unit and configuration validation, integrated process testing across order to cash and procure to pay, performance and volume testing on API-heavy interfaces, and a user acceptance wave with real transaction volumes. Cutover for a multi-site manufacturer typically consumes a long weekend with a rehearsed hour-by-hour runbook. What surprises most teams is life after go-live. CloudSuite delivers updates on a regular cadence rather than at your convenience, so you need a permanent, small regression suite covering critical flows, scripts, and interfaces, executed against each update in a non-production environment before it reaches you. Budget for that ongoing capacity in the operating model, not as project scope.

How Netray AI Agents De-Risk M3 Cloud Migration

Netray agents parse modification source, interface logs, and usage telemetry to produce a scored disposition register in days rather than the six to ten weeks a manual fit-gap normally takes, with effort estimates per item and a recommended cloud pattern for each. During the project, agents generate API-based replacements for retired direct database jobs and build the regression suite as a by-product of process testing rather than as a separate effort. After go-live those same tests run automatically against each CloudSuite update, so your team learns about a breaking change before your users do. For regulated manufacturers the whole toolchain can run inside your own network.

Frequently Asked Questions

How long does an Infor M3 CloudSuite migration take?

Most mid-size manufacturers plan nine to eighteen months from kickoff to go-live. Single-site organizations with few modifications land at the shorter end, while multi-site, multi-country groups with heavy customization and complex integrations extend beyond it. The dominant driver is the number of business engine modifications and direct database integrations that must be redesigned, not the technical migration itself. An honest inventory in the first six weeks is the best predictor of an accurate timeline.

Can you keep M3 modifications in CloudSuite multi-tenant?

No. Multi-tenant CloudSuite does not permit business engine modifications, custom code on the application servers, or direct database access. Every modification must be reclassified as configuration, H5 personalization or scripting, an API-based extension, an ION document flow, or an externally hosted service reacting to Event Hub events. Some organizations choose single-tenant hosting as an interim step, but multi-tenant is where the ongoing update cadence and lower operating cost live.

How much transaction history should we migrate to CloudSuite M3?

Migrate master data and all open items in full, then choose a defined window of closed transactions, commonly one to three fiscal years, for period-over-period comparability. Keep older history in a read-only archive or data lake that finance and audit can query rather than loading it into the live system. Full history migration multiplies conversion runtime and validation effort while adding little operational value, and it is the easiest scope reduction available to most projects.

Key Takeaways

  • 1Inventory Modifications, Interfaces, and Direct Database Access: Start with a complete register: every business engine modification, every custom program, every interface in or out, every report built on direct SQL, every scheduled script on the application server, and every H5 or Smart Office script. For each entry capture the business purpose, the owner, the last time it actually ran, and whether an equivalent capability now exists in standard M3.
  • 2Rebuilding Customizations as Cloud Extensibility: The cloud extensibility toolkit is genuinely capable once teams stop trying to reproduce a modification line for line. Field-level defaults, hidden fields, and guided flows become H5 personalization and scripting.
  • 3Data Migration and History Strategy: Deciding how much history to carry is the second largest cost driver. Full transactional history migration is slow, expensive to validate, and rarely justified.

Planning an M3 CloudSuite migration? Netray can deliver an AI-generated modification and interface disposition register in weeks, not months.