Distributed Cloud Configuration

The Distributed Cloud model in StarlingX enables centralized management and orchestration of a geographically distributed network of StarlingX clouds. It is designed to provide flexibility in managing cloud clusters and supporting various deployment architectures.

See Deployment Configuration Terminology for a description of common deployment configuration terminology, including definitions of different node types and network types.

• Architecture:

  A Distributed Cloud system consists of multiple clouds:

  • A large pool of geographically distributed autonomous clouds (i.e., subclouds), which run the latency or non-latency sensitive workloads at the edge. The deployment configuration of a subcloud can be any of the single-cloud deployment configurations supported (i.e., AIO-SX, AIO-DX with worker nodes, or Standard).

  • A central cloud (i.e., System Controller cloud) that provides automation and orchestration of lifecycle operations for the geographically distributed subclouds. The central cloud’s deployment configuration can be either AIO-DX with worker nodes or Standard.

• Typical Deployment:

  Ideal for central locations where workloads are geographically distributed and require efficient centralized management across multiple locations.

• Storage Options:

  The central cloud and subclouds can each have independent storage options, using either a local, internal Rook-managed Ceph cluster or an external storage backend (for example, NetApp Trident or a Dell Storage Array).

• Networking:

  The Central Cloud communicates with the subclouds over L3 networks using:

  • The OAM Interface

    and

  • The MGMT Interface (default) or the Admin Interface.

    • The admin network is recommended for new Distributed Cloud deployments to facilitate the separation of the system controller to subcloud traffic from the management network.

    • If IP Address changes are required, the Admin Interface should be used.