Large scale migrations

techhadoop azure

Migrating 6,000 Virtual Machines (VMs) to Azure moves you out of standard “lift-and-shift” operations and into large-scale cloud transformation. At this scale, manual or piecemeal approaches fail due to operational bottlenecks, bandwidth limitations, and business disruption.

The industry standard for large-scale migrations follows the Microsoft Cloud Adoption Framework (CAF), structured around four primary phases.

Architecture Phase Overview

 ┌──────────────────┐    ┌──────────────────┐    ┌──────────────────┐    ┌──────────────────┐
 │  1. DISCOVERY    │ ──►│   2. WAVE        │ ──►│   3. MIGRATION   │ ──►│  4. GOVERNANCE   │
 │   & ASSESSMENT   │    │    PLANNING      │    │    EXECUTION     │    │   & MONITORING   │
 └──────────────────┘    └──────────────────┘    └──────────────────┘    └──────────────────┘
   • Azure Migrate         • App Grouping          • Azure Migrate        • FinOps Cost Mgmt
   • RVTools Analysis      • Move Groups           • ExpressRoute Setup   • Defender for Cloud
   • TCO Modeling          • Network Mapping       • Automated Cutover    • Policy Enforcement

1. Discovery & Automated Assessment

At 6,000 VMs, automated tooling is essential to discover dependencies, right-size compute, and estimate total cost of ownership (TCO).

  • Deploy Azure Migrate Appliances: Deploy multiple agentless Azure Migrate appliances across your vSphere/Hyper-V infrastructure to continuously collect performance metrics (CPU, RAM, disk IOPS, network utilization).
  • Dependency Mapping: Use agentless or agent-based dependency analysis to group interconnected workloads. Never migrate a database without mapping its dependent application servers first.
  • Right-Sizing & Reserved Instances: Base sizing recommendations on performance-based data rather than static VM allocations. Calculate cost models using 3-Year Azure Reserved Instances (RIs) or Savings Plans, alongside Azure Hybrid Benefit (AHUB) for existing Windows Server and SQL licenses.

2. Landing Zone Foundations & Wave Planning

Before moving a single production workload, establish an enterprise-scale Azure Landing Zone.

                           ┌─────────────────────────────────┐
                           │      Tenant Root Group          │
                           └────────────────┬────────────────┘
                                            │
                               ┌────────────▼────────────┐
                               │    Management Group     │
                               └────────────┬────────────┘
                                            │
           ┌────────────────────────────────┼────────────────────────────────┐
           ▼                                ▼                                ▼
┌─────────────────────┐          ┌─────────────────────┐          ┌─────────────────────┐
│ Platform Management │          │   Identity & Auth   │          │ Connectivity (Hub)  │
│  • Azure Monitor    │          │ • Domain Controllers│          │  • ExpressRoute     │
│  • Log Analytics    │          │  • Entra ID Connect │          │  • Azure Firewall   │
└─────────────────────┘          └─────────────────────┘          └─────────────────────┘
Wave Planning & Categorization

Divide the 6,000 VMs into structured migration waves using the 6 Rs framework:

Migration StrategyTarget %Criteria
Rehost (Lift-and-Shift)60%–70%Legacy OS, monolithic apps, off-the-shelf software (ISV).
Replatform (Lift & Reshape)15%–20%Move SQL Server VMs to Azure SQL Managed Instance or Azure SQL DB.
Retire / Retain10%–15%Decommission unused VMs identified during discovery scans.
Refactor< 5%High-value internal apps converted to containers (AKS) or App Services.
Bandwidth & Rehydration Sizing

Transferring petabytes of data requires dedicated hybrid connectivity:

  • Connectivity: Establish dual Azure ExpressRoute circuits (10 Gbps) in an active-active, geographically redundant configuration.
  • Offline Data Transfer: For non-time-sensitive data stores, leverage Azure Data Box devices to perform initial seed transfers offline.

3. Migration Execution Engine

Executing a 6,000 VM migration requires automation and factory-style execution.

1.Establish Infrastructure & Waves:Prep Phase.

  1. Deploy Azure Landing Zone (Hub-and-Spoke network topology, ExpressRoute, Azure Policy).
  2. Define target wave sizes (typically 100–300 VMs per migration weekend).
  3. Validate DNS resolution and Active Directory replication between on-premises and Azure.

2.Initial & Continuous Data Sync:Replication Phase.

  1. Enable initial agentless block-level replication using Azure Migrate.
  2. Monitor bandwidth throttling across ExpressRoute to prevent impacting production on-premises workloads.
  3. Allow continuous delta sync to keep target Azure managed disks up to date prior to cutover.

3.Non-Disruptive Migration Rehearsal:Testing Phase.

  1. Perform Test Migrations on isolated, non-routed target Virtual Networks (VNets).
  2. Validate application functionality, service start behaviors, and database connection strings.
  3. Remediate missing drivers or network configuration issues before scheduling cutover.

4.Final Sync & System Cutover:Cutover Phase.

  1. Shut down source VMs on-premises to guarantee data quiescence.
  2. Trigger final delta synchronization in Azure Migrate.
  3. Provision and boot target Azure VMs, update internal DNS records, and run validation scripts.

4. Post-Migration Governance & Day-2 Operations

Once workloads arrive in Azure, shift immediately into operational governance:

  1. FinOps Cost Optimization:
    • Enable Azure Cost Management and set budget alerts at the Subscription and Resource Group levels.
    • Auto-shutdown non-production VMs outside business hours.
    • Apply Reserved Instances or Savings Plans immediately post-cutover.
  2. Security & Posture Management:
    • Onboard target subscriptions into Microsoft Defender for Cloud.
    • Enforce Azure Policy baselines (e.g., restrict public IPs, force encryption on managed disks, require specific tags).
  3. Backup & Disaster Recovery:
    • Enable Azure Backup with automated Vault policies.
    • Configure Azure Site Recovery (ASR) for critical workloads needing cross-region disaster recovery.

Leave a Reply