Skip to main content
cubecos
Version: 3.0

Get to Know CubeCOS Node Roles

Node Roles in CubeCOS​

Each CubeCOS cluster is made up of nodes with defined roles. These roles can be combined or separated depending on the deployment type and performance requirements.

The diagram below illustrates each type of node's distinct roles in CubeCOS.

node-roles-overview

Compute​

A Compute node provides processing power to run virtual machines, containers, and other workloads. It does not offer storage services directly, and instead relies on other nodes in the cluster for data storage.

  • Offers compute resources only
  • Connects to shared storage provided by storage-capable nodes
  • Ideal for scenarios where compute and storage scale independently

Compute with Storage Node​

This node type contributes both compute and storage resources to the cluster. It allows the platform to scale out horizontally, adding capacity where it’s needed most.

  • Adds CPU and memory for workloads
  • Provides additional storage capacity
  • Enables efficient scaling by combining functions in one node

Control Node​

Control nodes run the CubeCOS control plane, which includes the web-based dashboard and system management services (infrascope). In larger deployments, the control plane can be isolated to dedicated nodes for better performance and availability.

  • Hosts management and orchestration services
  • Improves cluster responsiveness in HCI and Web-Scale deployments
  • Recommended for clusters with high control plane traffic

Control-Converged Node​

A Control-Converged node combines compute, storage, network, and control plane functions into a single system. It is typically used in smaller clusters, such as All-In-One deployments, where simplicity and space efficiency are important.

  • Provides a complete hyperconverged stack in one node
  • Includes the CubeCOS dashboard and management services
  • Ideal for edge, branch, or SMB use cases

Storage Node​

Storage nodes expand the cluster’s capacity for persistent data. They are dedicated to handling high-throughput and high-capacity storage workloads, without contributing compute resources.

  • Adds disk capacity to the cluster
  • Supports storage-heavy workloads
  • Enables storage scaling without impacting compute