Canonical Ceph reference architecture¶
Canonical Ceph can be deployed as a stand-alone cluster, or integrated with OpenStack, VMware, or Kubernetes. Depending on your cluster use case, Canonical Ceph provides various deployment options: single-node and multi-node deployments, through MicroCeph; large-scale deployments through Charmed Ceph, and containerized deployments with Ceph Rocks. A single cluster can provide block, file, and object storage, and can be customized to meet price-performance requirements.
Our reference architecture provides detailed guidelines for optimizing your Ceph architecture. It starts by detailing the different deployment options and their architectures, then highlights the factors to consider when designing a Canonical Ceph cluster, and then outlines hardware specifications, including compute, storage, memory and networking requirements. The reference section of the reference architecture includes Canonical’s recommendations regarding hardware and software configurations, to further inform your architectural choices.
Warning
This reference architecture is a starting point, not a prescription. Before making any purchasing decisions, we recommend reviewing your plans with Canonical to ensure they align with your specific needs. Hardware requirements can vary significantly depending on your workload and goals.
Canonical is not liable or responsible for any equipment purchases made as a result of this reference architecture. Following these recommendations does not guarantee that the proposed hardware will meet the requirements of your project or use case.
To discuss your specific requirements, contact us.