VCF Automation 9.1: VM Apps vs All Apps Organizations, and How They Co-exist
VCF Automation 9.1 gives a provider two very different kinds of tenant: an Organization for VM Apps and an Organization for All Apps. They share a name and a console, but they consume infrastructure in completely different ways. That difference decides whether they can live on the same cluster, on separate clusters, or in another VCF instance.
This post explains:
What each organization type is and who it is for
The differences between them, side by side
Whether both can point to the same cluster
Whether they can use different clusters, workload domains or VCF instances
How to create each one, in the order that matters
Two organization types, two consumption models:
Organization for VM Apps:
This is the Aria Automation 8.x experience carried into VCF 9. It is built for teams that already automate VM-based applications and want to keep working the same way after the upgrade.
Self-contained: the organization owns its infrastructure definitions. The organization's cloud administrator adds cloud accounts (vCenter and NSX), then builds cloud zones, flavor and image mappings, network and storage profiles, projects and policies.
No Supervisor needed: it provisions directly onto vSphere clusters through its vCenter cloud accounts.
Cloud templates as code: teams design in the canvas and YAML editor, publish to the catalog, and extend with event subscriptions and VCF Operations orchestrator workflows.
Upgrades land here: an upgraded Aria Automation 8.x tenant becomes a VM Apps organization automatically.
Public clouds are winding down: from 9.1.1, native management of AWS, Azure and GCP resources in VM Apps organizations is deprecated and turned off by default.
Organization for All Apps:
This is the new VCF 9 cloud-consumption model, and Broadcom's recommended starting point for new environments.
Provider-allocated: the provider builds regions from Supervisor-enabled clusters and gives each organization a region quota. Tenants never see vCenter or NSX.
Supervisor required: workloads run in vSphere Namespaces on vSphere Supervisor.
VPC networking: organizations get NSX VPCs, subnets and transit gateways, with quota-based network management.
Many services: VMs, VKS Kubernetes clusters, volumes, databases, Harbor registries, secret store, DNS, certificates and AI workloads, consumed through blueprints, the catalog or the self-service console.
Built-in hierarchy: organization → region quota → project → namespace → workloads.
A Provider Consumption Organization is a special All Apps organization. Only one can be enabled. It adds an embedded orchestrator and lets the provider publish catalog items to other organizations.

VM Apps vs All Apps: the differences
Area | Organization for VM Apps | Organization for All Apps |
Who it's for | Existing Aria Automation users who want no change to their automation | New VCF 9 private cloud consumers; Broadcom's choice for greenfield |
Who owns infrastructure | The organization (cloud accounts, cloud zones) | The provider (regions, region quotas) |
Runs on | vSphere clusters via vCenter cloud accounts | vSphere Supervisor and vSphere Namespaces |
Networking | Existing VLAN (VDS) or NSX segments, via network profiles | NSX VPCs, subnets and transit gateways |
Region quota | Not used | Required before namespaces can be created |
Design and catalog | Cloud templates (canvas and YAML), catalog of imported templates | Blueprints, Content Hub, curated catalog and self-service console |
Kubernetes and services | VM-centric | VKS clusters, databases, registries, secrets and more |
Identity | Own identity source; upgraded tenants keep legacy authentication | Own identity provider per organization |
How many | One from the Provider Management UI; more only through the API | As many as you need |
Can VM Apps and All Apps co-exist?
Yes. Before 9.1, each organization type typically needed its own workload domain with its own vCenter and NSX. VCF 9.1 adds a bimodal design in which both organizations can share one vCenter and NSX. The answer depends on where you want them to run.

1. Same cluster
Supported, with conditions. Both organizations can deploy onto the same vSphere cluster when they share a vCenter and NSX. The All Apps side needs a Supervisor on that cluster, while VM Apps keeps provisioning to it directly.
Turn on the Mixed Tenancy Mode feature flag (Feature Flags in the Provider Management UI).
VM Apps workloads keep their existing VLAN or NSX segment networking. All Apps workloads use VPCs.
No strict isolation: VCF Automation doesn't enforce resource isolation between the two in this mode. Broadcom positions it for migration phases and trials, not for production environments that need strict multi-tenancy.
Noisy neighbours: one cluster costs less, but contention is likely. Keep it to small or pilot environments with similar workloads and relaxed SLAs.
2. Different clusters, same vCenter
Supported, and the better choice for larger environments. For example, Cluster A serves VM Apps and Cluster B runs the Supervisor for All Apps, both under the same vCenter and NSX. You get real performance isolation at the cost of more hardware, and Broadcom calls this mandatory for large deployments with mixed, mission-critical workloads. Because vCenter and NSX are still shared, the same Mixed Tenancy Mode flag and ordering rules apply.
3. Different workload domains, same VCF instance
Supported. This is the classic layout. The VM Apps organization uses cloud accounts for workload domain A. The All Apps region is built from a Supervisor in workload domain B. Nothing is shared at the vCenter or NSX level, so you get the strongest separation without relying on Mixed Tenancy Mode.
4. Another VCF instance
VM Apps: yes. A VM Apps organization can deploy and manage workloads across several workload domains and even several VCF fleets. Add a vCenter cloud account, with its NSX account, for each target. Broadcom documents this multi-fleet pattern for VM Apps only. It doesn't work with All Apps.
All Apps: yes, within the same fleet. One VCF Automation can connect to infrastructure in different VCF instances managed by the same VCF Operations. Create a separate region for each instance's Supervisors and give the organization a quota in each region.
All Apps across fleets: no. Each fleet runs its own VCF Automation, so a tenant in another fleet needs its own organization there.
Placement | VM Apps | All Apps | Notes |
Same cluster as the other org type | Yes | Yes | Mixed Tenancy Mode; no strict isolation; trials and migration |
Different cluster, same vCenter | Yes | Yes | Mixed Tenancy Mode; recommended for large deployments |
Different workload domain, same instance | Yes | Yes | No shared vCenter or NSX |
Different VCF instance, same fleet | Yes | Yes, through its own region | One NSX Manager per region |
Different VCF fleet | Yes | No | Multi-fleet pattern is VM Apps only |
Region rules that shape All Apps placement
For All Apps, regions decide where a tenant can run.
Building blocks: a region is made of one or more Supervisors, which can come from more than one vCenter.
One NSX per region: every vCenter in a region must use the same NSX Manager. Supervisors on vCenters with different NSX instances need separate regions.
Matching Supervisors: all Supervisors in a region need identically named storage and VM classes, the same Kubernetes version and the same IaaS services. From 9.1.1, they also need the same network stack type.
Quotas: an organization can hold quotas in several regions, but only one quota per region. The region of a quota can't be changed later.
Namespaces stay put: a namespace lives in one region.
Set them up in the right order
When VM Apps and All Apps share a vCenter and NSX Local Manager pair, order matters.
Turn on Mixed Tenancy Mode under Feature Flags.
Create the Organization for VM Apps first.
In that organization, add the vCenter and NSX cloud accounts and finish its configuration.
Only then create the region using the same vCenter and NSX Manager.
Create the Organization for All Apps and give it a quota in that region.

Create the Organization for VM Apps
In the Provider Management UI, go to Feature Flags and make sure VM Apps Organization Creation is active.
Go to Infrastructure → Organizations, click Create Organization, choose Organization for VM Apps and click Next.
Enter a name (it becomes part of the organization's URL) and an optional description. Click Create and Continue. From here, creation can't be cancelled; to remove the organization later, deactivate it, then delete it.
Add the first user with a password of at least 15 characters. Give it Organization Owner, plus Assembler Administrator, Orchestrator Administrator and Service Broker Administrator, so you keep full access if the identity provider is unreachable.
Click Add User and Finish. The first user then signs in and connects the organization to your identity provider.
Inside the organization, add the vCenter cloud account under Infrastructure → Connections → Cloud Accounts → Add Cloud Account. Choose at least one data center and link the NSX cloud account. The NSX-to-vCenter association itself is configured in NSX, not in VCF Automation.
Create the Organization for All Apps
Make sure a region exists, built from a Supervisor-enabled vCenter.
Go to Infrastructure → Organizations → Create Organization. If a VM Apps organization already exists, the wizard goes straight to the All Apps flow.
Enter the name and click Create and Continue.
In Assign a Region Quota, pick the region. Then share either the whole region or a portion of it, with per-zone CPU and memory limits and reservations, VM classes and storage classes. The region can't be changed afterwards.
In Configure Networking, choose Centralized Connection, Distributed VLAN Connection or Distributed VXLAN Connection, depending on what the region offers, and pick the external connection.
Add the first user with the Organization Administrator role, then click Add User and Finish.
Which should you use?
Upgrading from Aria Automation 8.18? Your tenant becomes a VM Apps organization. Keep it running, add an All Apps organization next to it, then pilot and move workloads across over time.
Starting fresh? Begin with All Apps.
Need VMs in another fleet or many workload domains? That's a VM Apps pattern.
Production with strict tenant isolation? Don't rely on Mixed Tenancy Mode. Use separate workload domains, or at least separate vCenter and NSX boundaries.



Comments