VMware migration to Azure and Kubernetes with agentic discovery, dependency-aware migration waves, Proxmox and OpenShift alternatives, Terraform/Bicep, Ansible, FinOps and disaster-recovery gates.
VMware exit and application modernization are not destination-selection exercises. They are dependency, continuity, licensing, network, recovery and unit-economics programs. This repository implements a deterministic decision factory that turns an estate into dispositions, destination recommendations, migration waves and an auditable business case.
Evidence status: the Python engine, tests and 100-VM synthetic case are implemented. Azure, VMware, Proxmox, OpenShift, Kubernetes and Ansible artifacts are integration contracts until exercised in an authorized environment.
For each workload, should we retain, retire, rehost or replatform it;
where can it move without violating dependencies and support constraints;
in which wave; with what rollback; and at what fully loaded cost?
- Validates asset dependencies and rejects missing nodes and cycles.
- Calculates transitive blast radius.
- Recommends retain, retire, rehost or replatform.
- Rejects destinations that violate declared legacy or container requirements.
- Selects the lowest modeled cost among compatible destinations.
- Requires approval for critical, shared and retired workloads.
- Generates dependency-safe migration waves.
- Includes current, destination-shared and per-workload costs.
- Calculates annual run-rate delta, three-year net benefit, ROI and payback.
- Emits a simulated evidence receipt with an explicit integrity disclaimer.
flowchart LR
INV[RVTools / CMDB / APIs / Terraform state] --> NORM[Normalized estate]
FLOW[Network and dependency evidence] --> GRAPH[Dependency graph]
NORM --> GRAPH
GRAPH --> DISP[Disposition engine]
DISP --> OPT[Destination and TCO optimizer]
OPT --> WAVE[Migration wave planner]
WAVE --> GEN[IaC / Ansible / test / rollback contracts]
GEN --> GATE[Policy + owner approval]
GATE --> CUT[Cutover and validation]
CUT --> REC[Cost, performance and recovery reconciliation]
REC --> EVID[Evidence receipt]
Read architecture.md for execution and identity boundaries.
No third-party Python dependency is required.
python3 examples/generate_estate.py
PYTHONPATH=src python3 -m unittest discover -s tests -v
PYTHONPATH=src python3 -m modernize.cli \
examples/synthetic-100-vm-estate.json \
--output evidence/synthetic-analysis.jsonThe fixture contains 100 synthetic workloads across databases, ERP, APIs, identity, monitoring, file, web, batch, worker and legacy roles. It includes production/nonproduction separation, criticality, utilization, data classes, dependencies, compatibility and three end-of-life workloads.
| Metric | Synthetic result |
|---|---|
| Workloads | 100 |
| Dependency-safe waves | 2 |
| Current annual cost | $1,360,000 |
| Destination annual cost | approximately $621,010 |
| Migration investment | $650,000 |
| Annual run-rate delta | approximately $738,990 |
| Modeled payback | 10.55 months |
These numbers are sensitivity-model outputs, not customer observations, quotes or guaranteed savings. See unit economics and assumptions.
The generic schema supports modeling:
- Azure Virtual Machines, Azure VMware Solution, Azure Local and AKS;
- AWS EC2/EKS and GCP Compute Engine/GKE;
- VMware Cloud Foundation, Hyper-V and Nutanix;
- Proxmox VE, KVM, OpenShift Virtualization/KubeVirt;
- Apache CloudStack, OpenStack, Kubernetes and XCP-ng.
The fixture directly compares generic Azure-native, AKS, Proxmox and OpenShift Virtualization coefficients. Named product adapters remain contracts.
- Azure Bicep: migration network, Recovery Services vault and Log Analytics.
- Ansible validation: OS, application port, filesystem and rollback-reference checks.
- Migration policy: cutover prerequisites, human approvals and prohibited automation.
- CI: tests, fixture regeneration and end-to-end analysis.
- Missing discovery evidence becomes a blocker, never an inferred fact.
- LLM-generated recommendations cannot execute production changes.
- Discovery and execution identities must remain separate.
- Critical and shared workloads require owner approval and parallel rollback.
- Retirement always requires approval.
- A snapshot is not represented as a database consistency guarantee.
- The optimizer may recommend retaining VMware when no safe destination exists.
- Implemented: executable source and automated tests.
- Simulated: deterministic synthetic estate and modeled economics.
- Contract: provider, automation or IaC boundary not exercised here.
- Deployed: reserved for evidence captured from an authorized environment.
The SHA-256 receipt detects serialized-envelope changes; it does not prove inventory completeness, identity, custody, migration or realized savings.
| Domain | KPIs |
|---|---|
| Discovery | coverage, unknown owner rate, unclassified traffic, stale assets |
| Planning | recommendation confidence, blocked workloads, wave readiness |
| Delivery | migration throughput, schedule variance, change failure rate |
| Reliability | downtime, rollback rate, recovery validation, RTO/RPO attainment |
| Economics | cost per migrated workload, run-rate delta, payback, variance |
| Modernization | retired, rehosted, replatformed and containerized percentages |
| Operations | utilization, incidents, support effort and platform SLOs |
- Estate and dependency assessment.
- Executive destination and unit-economics decision.
- Pilot migration wave with rollback evidence.
- Factory-scale migration and modernization.
- Managed hybrid platform, FinOps, reliability and compliance.
Pricing must follow estate size, dependency complexity, data volume, criticality and accountability—never VM count alone.
VMware migration, VMware alternatives, VMware to Azure migration, cloud migration, application modernization, Azure migration, Kubernetes migration, Proxmox, OpenShift Virtualization, KubeVirt, Terraform, infrastructure as code, Ansible automation, hybrid cloud, multicloud, FinOps, disaster recovery, cloud cost optimization, platform engineering, Principal Cloud Architect and Forward Deployed Engineer.
Planning a VMware renewal, data-center exit or application-modernization program?
Request a migration architecture and unit-economics assessment.