Jump to: navigation, search

StarlingX/Docs and Infra/Greg/configurationOptions

Configuration Options

The StarlingX configuration options are:

  • Single all-in-one server
  • Two HA all-in-one servers
  • Two HA controller servers and two compute servers
  • Two HA controller servers and two compute servers with two HA storage server cluster


Single all-in-one server

Starlingx-deployment-options-simplex1.png
The simplex configuration runs all Cloud functions (i.e. control, storage, and application workloads) on one node. This configuration is intended for very small and physically isolated Edge sites that either do not require overall server hardware redundancy or it is done at a higher level. Hardware component protection could be used in this solution if, for example, an HW RAID or 2x Port LAG is used in the deployment. Simplex is a good configuration to use when evaluating StarlingX.

Two HA all-in-one servers

Starlingx-deployment-options-duplex1.png
The duplex configuration runs all Cloud functions (i.e. control, storage, and application workloads) on one node, but a second node exists for Active / Standby based hardware redundancy for all platform and optionally application services. Duplex is a good configuration to use when evaluating StarlingX’s High Availability capabilities.

Two HA controller servers and two compute servers

Starlingx-deployment-options-controller-storage1.png
This configuration provides a two node HA Controller cluster that also provide small-scale HA Storage for the Cloud. The configuration allows for between one and approximately 10 Compute nodes to run application workloads. The restriction on the size of the compute node pool is due to the integrated control and storage services in the controller cluster. This configuration works best for Clouds with smaller capacity needs.

Two HA controller servers and two compute servers with two HA storage server cluster

Starlingx-deployment-options-dedicated-storage1.png
This configuration is similar to the “Standard with controller storage” configuration but has dedicated and more scalable Storage node cluster in addition to the Controller and Compute nodes. The larger scale and non-integrated storage cluster allows scaling the compute node pool up to 100 compute nodes. You can use this configuration for Clouds that require large scale application workloads.