Broadberry Ceph Storage

One Platform. Three Storage Services.

Ceph provides a common distributed storage foundation while allowing applications to access data through the interface they require.

These are separate access models. Data stored through one service does not automatically appear through another.

Ceph Architecture Services

Before recommending a Ceph architecture, Broadberry evaluates your workloads, application services, capacity, performance and resilience requirements.

Choose the Right Ceph Architecture for Your Workload

Ceph can support very different storage requirements. The application determines the service, media and architecture.

Flexible Performance and Capacity

Ceph lets Broadberry tailor storage media, data protection and infrastructure to different workload requirements.

Capacity + Performance Pools

Combine HDD capacity with SSD or NVMe performance pools.

Distributed Protection

Protect data across drives, hosts and racks using defined failure domains.

Scalable Architecture

Add OSD nodes as capacity and workload requirements grow.

Open Hardware

Deploy Ceph on standard, non-proprietary server infrastructure.

Design for Usable Capacity, Not Raw Capacity

Usable capacity depends on protection, failure domains and recovery requirements. Broadberry designs Ceph storage to deliver the capacity your applications can actually use.

Ceph Storage Use Cases

Ceph supports a wide range of workloads across file, block and object storage.

Why Broadberry for Ceph?

Broadberry combines Ceph expertise with workload-led hardware design to deliver a complete, validated storage platform rather than simply supplying the servers.

Workload-Led Design

Designed around your applications, performance and capacity needs.

Complete Cluster Engineering

Compute, memory, media, networking and Ceph services designed together.

Flexible Hardware

Choose HDD, SSD and NVMe configurations.

Deployment and Lifecycle Support

Support from planning through future expansion.

Frequently Asked Questions About Ceph Storage

What is the difference between Ceph and CephFS?

Ceph is the distributed storage platform. CephFS is its shared filesystem. Ceph can also provide block storage through RBD and object storage through RGW.

Can Ceph support virtual machines?

Yes. RBD provides block devices through supported host and hypervisor integrations. The configuration should be sized and validated for the expected VM workload and recovery requirements.

Can Ceph use both HDD and NVMe storage?

Yes. Different media classes and placement rules can be used to create capacity and performance pools. Adding NVMe, however, does not automatically accelerate all data in the cluster.

Is Ceph suitable for AI and GPU clusters?

Yes, when the storage media, metadata services and networking are designed to meet the application's performance requirements. CephFS can provide shared file access across GPU hosts, while flash pools can be used for active training data.

Should every Ceph pool use erasure coding?

No. Replication and erasure coding have different capacity, performance and recovery trade-offs. CephFS metadata pools require replication, while appropriate data pools may use erasure coding.

Does Ceph eliminate the need for backup?

No. Distributed redundancy and snapshots protect against different risks than an independent backup. Backup, retention and off-site recovery should be planned separately.

Can Ceph scale as storage requirements grow?

Yes. Additional OSD nodes can add capacity and resources, but expansion triggers data placement changes and rebalancing. Broadberry plans capacity reserve, networking and recovery requirements as part of the architecture.

Build the Right Ceph Storage Platform

Tell us about your applications, capacity, performance and recovery requirements. Broadberry will design a Ceph storage architecture around the services, media and protection policies your environment needs.