Discover
Normalize mixed estates into one inventory model.
Hypervisor-agnostic workload migration
Viaduct gives operators one system for discovery, dependency-aware planning, migration orchestration, rollback and resume awareness, and post-cutover lifecycle visibility across mixed hypervisor estates.
Discover
Normalize mixed estates into one inventory model.
Map
Model workload, storage, network, and backup dependencies.
Migrate
Use checkpoints, approvals, rollback, and resume-aware execution.
Operate
Track lifecycle posture, drift, policy, and cost after cutover.
Workflow
Viaduct keeps the migration path legible: normalize the estate, model dependencies, validate cutover, execute with checkpoints, and keep lifecycle signals visible after the move.
Normalize mixed estates
Model dependencies
Validate cutover
Checkpoint and resume
Track drift and cost
Architecture
Connectors feed a shared system of record, the core engines add dependency and execution logic, and the CLI, API, and dashboard all reflect the same persisted truth.
Hypervisor and backup adapters with a restrained plugin model.
A staged pipeline with shared context instead of siloed utilities.
Concurrent, read-only collection normalizes inventory and infrastructure metadata into the universal model.
Cross-platform node and edge modeling exposes workload, storage, network, and backup relationships before execution.
Declarative plans drive validation, wave execution, checkpoints, cutover, rollback, and resume behavior.
Post-migration analysis keeps drift, cost, policy, portability, and remediation visible in the same control plane.
Operator surfaces that stay in parity because they read and write the same state.
Declarative specs, discovery, planning, status, and rollback.
Tenant-scoped routes for inventory, graphs, migration state, policy, reporting, and metrics.
Inventory, dependency, migration, lifecycle, and reporting views on one backend model.
Execution context stays durable so retries, approvals, and rollback do not lose the thread.
Baseline inventory, infrastructure, and historical diffs.
Execution plans, approvals, checkpoints, and resume metadata.
Rollback artifacts and portability cleanup context.
Quotas, service accounts, permissions, and request-scoped context.
Execution model
Read-only discovery across mixed source platforms
Pre-flight validation, wave planning, approvals, and checkpoints before cutover
Lifecycle, policy, cost, and remediation remain visible after migration completes
Platform coverage
The platform row stays technical and quiet on purpose. Connector availability matters, but the real story is the shared migration and lifecycle control plane behind it.
vSphere inventory and migration planning
REST-backed discovery and target planning
WinRM and PowerShell-driven collection
Libvirt-ready with portable fallback support
Prism-backed mixed-estate visibility
Backup correlation and portability context
gRPC extension path with compatibility checks
Start in a lab
The first-run path starts with the local KVM lab under `examples/lab/`, then expands into live discovery, dependency analysis, migration pilots, and dashboard-backed operations.