Raff Volumes are NVMe block storage disks from 10 GiB to 1 TB. Attach one to any VM, grow it online without a reboot, snapshot it anytime, and keep it when the server goes away. Priced at $0.08 per GiB-month, traffic between VM and volume free and unmetered.
Servers are disposable. Data isn't. A volume separates the two, so you can rebuild, resize or replace the VM without ever touching what's on the disk.
One storage class, one price, no performance tiers to decode. Every volume gets the full feature set.
Every volume runs on NVMe SSD storage. Start at 10 GiB and grow to 1 TB as you need it.
Grow a volume while it is attached and in use. No reboot, no maintenance window.
Destroy the server and the volume detaches instead of dying. Attach it to the next VM.
Point-in-time snapshots of any volume, billed only above your free storage pool.
Traffic between your VM and its volumes is never metered and never billed.
Volume data is encrypted at rest on the platform by default. Nothing to configure.
Linux volumes arrive formatted ext4 or xfs and mounted at /dev/vdb. Windows formats NTFS in one click.
Volumes follow your IAM projects: volume.create and volume.delete are separate permissions.
Full create, attach, detach and resize in the REST API, the raff CLI and the Terraform provider.
No IOPS tiers, no provisioned-throughput math, no per-request fees. The price below is read live from our pricing database.
Yearly billing: $0.96/GiB-year (2 months free). Snapshots share your free backup-storage pool; I/O between VM and volume is free.
Where each provider lands on price and the fine print. AWS wins on exotic performance tiers; most workloads never need them.
| Raff VolumesBest | AWS EBS (gp3) | DigitalOcean | |
|---|---|---|---|
| Price per GB-month | $0.08 | $0.08 | $0.10 |
| Storage class | NVMe, one class | Tiered (gp3/io2/st1…) | SSD |
| Provisioned IOPS math | None to do | Baseline + paid IOPS/throughput | None |
| Online grow | Yes, attached and live | Yes | Yes (resize, reboot for some) |
| Survives server deletion | Yes, auto-detach | Depends on flag | Yes |
| Snapshot billing | Free pool, then $0.06/GB | $0.05/GB-mo | $0.06/GiB-mo |
| Volume I/O charges | None | None (gp3 base) | None |
Raff price is read live from our pricing database. Competitor rows reflect published us-east pricing and docs, gathered Jul 2026. Re-verify before relying on them.
No LUN mapping, no multipath config, no storage tickets. Three steps and you have durable storage.
Pick a size from 10 GiB to 1 TB and a VM. The volume attaches as /dev/vdb in seconds.
Linux volumes arrive formatted ext4 or xfs and mounted. Windows formats NTFS from Disk Management.
Resize online from the dashboard, API or Terraform. The filesystem expands without a reboot.
Volumes are the durable layer under the rest of Raff. Everything below runs on the same storage fabric.
Block storage is a raw disk you attach to a server, formatted with a filesystem like any drive. On Raff, volumes are NVMe disks from 10 GiB to 1 TB that attach to your VMs in seconds and keep existing when the VM is deleted.
Volumes are billed per GiB per month at the rate shown on the volumes page, read live from our pricing system. There are no IOPS tiers, no throughput surcharges and no fees for traffic between the VM and its volume.
Yes. Volumes grow online while attached and in use: pick a new size and the disk expands with no reboot. Growing is instant; shrinking is not supported, so start small and grow when needed.
Nothing bad. The volume automatically detaches and stays in your account with its data intact. Attach it to another VM whenever you like; you only pay for the volume itself.
As a standard virtio disk: /dev/vdb, /dev/vdc and so on. Linux volumes arrive formatted ext4 or xfs and mounted; on Windows you format NTFS from Disk Management.
Yes. Volume snapshots are point-in-time copies you can restore from, and they share your account's free backup-storage pool. Only snapshot storage above the pool is billed per GB.
Use a volume when software expects a disk: databases, file systems, application data. Use object storage for files accessed over an API: media, backups, static assets. On Raff they cost and behave differently, so most stacks use both.
NVMe volumes that attach in seconds, grow online and outlive any VM. 15,000+ builders run on Raff in our us-east region, backed by a 99.9% uptime SLA in writing.
Guides to help you plan, build, and scale on Raff.
Series
Help teams design production storage and recovery across VM disks, block volumes, object storage, snapshots, database-aware backups, retention, restore testing, and disaster recovery.
Series
Container Networking & StorageCover Kubernetes-specific networking and persistent storage, including Services, ingress, DNS, network policies, volumes, storage classes, and workload recovery boundaries.