Skip to content
Live $300 credit for new accounts Valid for 60 days from account creation Get started →

Overview

ZSoftly Cloud Storage (ZCS) is a standalone, storage-only product. You get a dedicated, single-tenant Ceph storage cluster with root-level access, sized to your workload. It is not a compute cloud. It is the storage tier you build large-scale object, block, and file workloads on, on its own or alongside ZSoftly Public Cloud (ZCP) or Private Cloud (ZPCP).

A fully distributed Ceph cluster, provisioned on dedicated hardware and handed to you with administrative access. There is no single point of failure and no multi-tenant contention, so you get predictable performance and the freedom to tune the cluster to your workload. Pool layouts, replication versus erasure coding, device classes, and caching are all yours to configure.

A single cluster speaks three storage interfaces at once:

  • Object storage: an S3-compatible endpoint (Ceph RGW) for the S3 tools and SDKs you already use. See Object Storage.
  • Block storage: network block devices (Ceph RBD) for virtual machines, databases, and Kubernetes persistent volumes. See Block Storage.
  • File storage: a POSIX shared filesystem (CephFS) you mount across many clients. See File Storage.

Hosted (public) ZSoftly provisions and operates your dedicated cluster on ZSoftly hardware in one of our regions (YUL, YOW). You get root access and a private endpoint without owning or racking hardware.

On-premises (private) ZSoftly designs and deploys a dedicated cluster on hardware you own or operate, in your facility. This fits data residency, sovereignty, or existing data-center investments.

In both models the cluster is single-tenant and yours to administer. Same Ceph, same root access, same interfaces. Only the location and ownership of the hardware differ.

  • Exabyte scale, no single point of failure. Scale horizontally by adding nodes. Data is replicated or erasure-coded across the cluster.
  • Root-level control. Administer pools, users, OSDs, and the Ceph Dashboard directly.
  • Performance tiering. Use NVMe as a fast pool or a cache layer in front of high-capacity drives. See Performance & Tiering.
  • Replication and disaster recovery. Replicate to another ZSoftly region, an on-prem cluster, or any Ceph cluster you administer. See Replication & DR.
  • Version choice. Provisioned on a recent stable Ceph release, or a specific version you select with our team. See Ceph Versions.
  • Big data and analytics. Large datasets with parallel access and built-in redundancy.
  • HPC, ML, and AI. High-throughput shared storage for simulations and training datasets.
  • S3 at scale. An open, cost-effective alternative to hyperscaler object storage.
  • Backup and disaster recovery. A replication target for an existing Ceph cluster.

ZCS is dedicated and root-access. ZSoftly Public Cloud also offers object storage and block storage, but those are managed, multi-tenant services you consume through the ZCP portal and API. You use the storage, you do not operate the cluster. Choose ZCP storage for managed, pay-as-you-go capacity inside the public cloud. Choose ZCS when you need a dedicated cluster you control.

Contact ZSoftly to size and provision a cluster.

Once your cluster is provisioned, continue to Provisioning a Cluster and Accessing Your Cluster.