Cloud Modernization Patterns Module 1 · Size It Up

What Modernization Actually Means

Last reviewed · content updated

Beginner

What you'll learn

~12 min
  • Map a realistic legacy estate: a low-code app, a .NET Framework monolith, and a decade of file-share data
  • Apply the six Rs (rehost, replatform, refactor, rebuild, replace, retire) to real workloads
  • Explain why modernization means reducing what you operate, not moving where you operate it

Welcome to Meridian Utilities

You just started as the modernization architect at Meridian Utilities, a fictional regional water-and-electric utility with about 1,400 employees. On your first day, the IT director walks you through the estate:

  • FieldDesk — a service-request app that an operations analyst named Dana built on a low-code platform six years ago. It was supposed to be a stopgap. Today, 300 field technicians run their entire day through it, someone bolted SLA tracking onto it, and it exchanges data with two other systems via emailed CSV exports. Nobody is allowed to touch it on storm days.
  • AssetTrack — a fifteen-year-old .NET Framework 4.8 monolith that tracks every transformer, pump, and substation Meridian owns. It runs on two Windows Server 2016 VMs behind IIS, talks to an Oracle 19c database, and runs nightly batch jobs that three departments depend on.
  • The reporting share — ten years of nightly CSV extracts on a Windows file share, feeding a jungle of Excel workbooks and Access databases that management calls “the reports.”

The CIO’s direction is one sentence: “Get us out of the VM business without breaking the utility.”

This training follows that journey end to end. Every artifact you produce along the way — assessments, decision matrices, Dockerfiles, migration plans, test matrices — you will build with your AI CLI doing the heavy lifting.

The six Rs, in one pass

Every workload in a modernization gets one of six dispositions. Here is the vocabulary, applied immediately to Meridian:

DispositionWhat it meansMeridian candidate
Rehost (“lift and shift”)Move the VM as-is to cloud infrastructureAssetTrack’s VMs, as a first step only
ReplatformSmall changes to run on a managed serviceAssetTrack’s batch jobs onto a scheduler service
Refactor / RearchitectRestructure the code for cloud-native operationAssetTrack’s API, ported to modern .NET in a container
RebuildRewrite from scratch on a modern stackFieldDesk, if the analysis supports it
ReplaceBuy or adopt something that already existsFieldDesk, if an off-the-shelf product fits
RetireTurn it off and archiveHalf the Access databases nobody has opened since 2021

Naming note: the canonical cloud-migration “6 Rs” (Gartner/AWS) run rehost, replatform, repurchase, refactor, retire, retain — where retain means deliberately keeping a workload where it is, and buying something falls under repurchase. This training folds rebuild/replace together for Meridian’s story; when you meet the other list in the wild, it is the same conversation.

Notice something: FieldDesk appears twice with a question mark, and AssetTrack’s answer is “several Rs at once, in sequence.” That is normal. Real systems decompose into parts with different dispositions — and the disposition you pick before assessing is usually wrong.

💬Lift-and-shift is a tactic, not a strategy

Rehosting AssetTrack’s two VMs into the cloud changes your data-center lease, not your life. You still patch Windows, you still babysit IIS, you still own every 2 a.m. failure. Rehost buys you a landing zone and an exit from the server room — it does not reduce what you operate. The reduction comes from the later Rs, which is why this training spends most of its time there.

The real goal: operate less

Here is the framing that will carry through this whole training: modernization is the systematic transfer of operational burden to someone whose business is carrying it.

  • A VM makes you responsible for the OS, the runtime, the patching, the scaling, and the app.
  • A managed platform service makes the cloud provider responsible for everything below your code.
  • Every rung between those two is a trade: control you keep versus operations you shed.

When you evaluate any modernization decision in this training, the first question is never “what is the newest technology?” It is: “what is the most someone else can operate for me while still meeting this workload’s requirements?” In Lesson 2.1 that question becomes a formal decision ladder — and you will see that Kubernetes, which many teams treat as the default destination, sits on the rung you should reach for last.

Where AI fits

A modernization program is mostly reading and writing: reading old code, old configs, old tickets; writing assessments, decision documents, migration plans, and new code. That is exactly the work AI CLI tools compress. Across this training you will:

  • Point your AI CLI at a legacy codebase and get a structured risk assessment in minutes (Lesson 1.2)
  • Have it argue three competing futures for FieldDesk, each on its own terms (Lesson 1.3)
  • Generate and then adversarially critique deployment artifacts (Lessons 2.2-2.5)
  • Design a data migration with exact seams and an audit trail (Module 3)

The AI does not make the decisions. You do. The AI makes every decision better informed and every artifact faster to produce.

KNOWLEDGE CHECK

Meridian rehosts AssetTrack's two Windows VMs into the cloud unchanged. Which statement best describes what changed?

Key takeaway

Modernization is not a destination platform — it is the deliberate reduction of what you operate. The six Rs are your vocabulary, per-component dispositions are your unit of decision, and the assessment you are about to run in Lesson 1.2 is what turns opinions about Meridian’s estate into evidence.

Search lessons