This section provides detailed technical information on the supported deployment configurations for StarlingX. It focuses on the architecture, node roles, network options, storage options, redundancy, and scalability limits of each deployment configuration.

This section complements the StarlingX Introduction, which serves as a foundational resource for understanding the StarlingX solution. While the StarlingX Introduction provides a broad overview of the platform, this section offers detailed, deployment-specific information. For comprehensive planning considerations, such as hardware requirements, network design, and deployment strategies, refer to the StarlingX Planning.

This content is intended for users involved in:

• Designing infrastructure for edge or data center deployments.

• Planning and implementing StarlingX installations.

Users are expected to have a working knowledge of Kubernetes, cloud-native infrastructure, and basic networking concepts.

Use this Deployment Configuration section to:

• Understand the different multi-node system architectures of the different deployment configurations

  • the different node types and their roles

  • the different networks and their purpose

  • and the redundancy and scalability of the different node type groups

• Understand the networking solutions within each deployment configuration.

• Understand the storage solutions within each deployment configuration.

• Compare the different deployment configurations based on scalability and redundancy.

• Identify the right deployment configuration using the comparison matrix.

All deployments consist of a combination of controller, worker, and storage hosts. Controller hosts run the Kubernetes control plane. Worker hosts run Kubernetes application workloads. Storage hosts are not part of the Kubernetes cluster and are used to establish a dedicated bare metal Ceph cluster for Kubernetes PVC. Controllers can be a standard or All-in-One configuration, where All-in-One controllers also run Kubernetes application workloads in addition to the Kubernetes control plane. All hosts can be configured for standard or low-latency profiles based on kernel selection.

Note

Rook-Ceph is recommended for new deployments, as some host-based Ceph deployments may not be upgradable.

• Deployment Configurations
• Deployment Configuration Terminology
• All-in-one Simplex (AIO-SX) Configuration
• All-in-one Duplex (AIO-DX) Configuration
• Standard Configuration
• Distributed Cloud Configuration
• Deployment Model Comparison Matrix
• Worker Function Performance Profiles

• Deploy Edgeworker Nodes