Deployment models¶
An RG deployment is defined by three orthogonal dimensions that combine freely: hosting (cloud-hosted or on-premises), connectivity (connected or air-gapped), and tenancy (shared or private). Any combination is valid, and one organization can run several combinations at once under the same identities and permission model. The dimensions are independent — choosing on-premises does not force air-gapped, and choosing private does not force on-premises.
The three dimensions¶
Each dimension answers a different question and can be set independently of the others.
| Dimension | Options | What it decides |
|---|---|---|
| Hosting | cloud-hosted / on-premises | Where the cluster runs — on RemoteGenius-provided infrastructure or inside the operator's own environment. |
| Connectivity | connected / air-gapped | Whether the cluster reaches the cloud layer at all, or operates with zero external connectivity. |
| Tenancy | shared / private | Whether a cluster hosts multiple tenants with enforced isolation, or is dedicated to one. |
Because the dimensions are orthogonal, they multiply rather than trade off: an on-premises cluster can be connected or air-gapped, and shared or private; a cloud-hosted cluster is connected but can still be shared or private. The operator picks each dimension for its own reason — data residency drives hosting, security posture drives connectivity, and customer separation drives tenancy.
Reading the matrix¶
The combinations that matter in practice follow directly from the dimensions. A cloud-hosted, connected, shared cluster is the default multi-tenant deployment on RemoteGenius-provided infrastructure. A cloud-hosted, connected, private cluster dedicates that infrastructure to one tenant. An on-premises, connected, private cluster runs inside the operator's environment but still replicates audit upward and uses the cloud for sign-in. An on-premises, air-gapped, private cluster runs entirely within a controlled boundary with no external connectivity at all. Every combination shares the same cluster architecture and the same two-tier RBAC — the dimensions change where the cluster runs and what it talks to, not how devices are managed or how access is governed.
One organization, several clusters¶
The dimensions being orthogonal means a single organization can operate a mix of deployments simultaneously, all under one set of identities and one permission model. Consider one customer running three clusters at once: a connected cloud cluster for its general fleet, an on-premises connected cluster at a facility with data-residency requirements, and an air-gapped cluster at a secured site with no external connectivity. The same users, organization-scope roles, and per-cluster overrides apply across all three; a user's authority in the air-gapped cluster is governed by the same RBAC model as in the cloud cluster, evaluated locally there. The organization sees one inventory and one governance model spanning deployments that differ in hosting, connectivity, and tenancy.