High-performance block storage and S3-compatible object storage with predictable latency and clear limits.
Attached to VM · From €0.03/GB/mo
Software-defined storage volumes attached directly to your VM. Optimized for low latency and consistent IOPS.
Multi-protocol · €0.02/GB/mo
S3-compatible object storage with FTP, SFTP, and WebDAV access. Same files, different protocols.
Block and Object storage are built on the same software-defined storage layer, but optimized for different access patterns.
Predictable provisioning, no surprises
Volume creation request accepted.
~150msSpace reserved on storage nodes. Replication targets selected.
~500msAll replicas initialized. Volume available for attachment.
~1.2sVolume attached to VM. I/O ready.
~300msTotal time from API call to I/O-ready: under 3 seconds. No lazy allocation. No surprise latency spikes.
Choose your performance tier. SDS for durability, ephemeral for raw speed.
Dev, test, archives
Small DBs, CMS
Production DBs, APIs
High-traffic, analytics
OLTP, time-series
Caches, temp, scratch
No replication. No guarantees. Data lost on host failure.
SDS classes: sustained guarantees, not burst maximums. Ephemeral: best-effort local NVMe.
Not sure? standard works great for most production workloads. Use essential for dev/staging, ephemeral only for data you can recreate.
Quick guide based on your use case
Development environments, test databases, staging servers.
PostgreSQL, MySQL, MongoDB with real users.
OLTP, time-series, analytics, high-frequency writes.
Redis, Memcached, build artifacts, scratch space.
Not sure? Start with standard — it handles 90% of production workloads. Upgrade to enhanced if you see I/O bottlenecks. See full pricing →
How storage performance actually works
Each volume has a hard IOPS and throughput cap based on its class. These limits are enforced at the storage layer, not at the hypervisor.
We don't use a credit system. Your IOPS limit is sustained 24/7. No "baseline" that throttles after burst.
IOPS limits are enforced per volume. Your neighbors can't saturate your storage path. Each volume gets dedicated bandwidth to storage nodes.
Local NVMe on the host. No IOPS limits, best-effort performance. Data lost if host fails. Use for caches, temp files, scratch space.
€0.03/GB/mo. Same price as value SDS, but no durability guarantees.No "baseline IOPS" that throttles after you exhaust credits
No fighting for bandwidth with other tenants on the same controller
Space is reserved upfront. No surprise latency when you first write
IOPS specs are sustained guarantees, not marketing maximums
API-first storage for files, backups, and static assets
Same files, different protocols. Access your buckets via S3 API, FTP, SFTP, or WebDAV. Use the protocol that fits your workflow. All pointing to the same data.
Works with AWS SDK, s3cmd, rclone, and any S3-compatible client. Drop-in replacement for most S3 use cases.
Strong read-after-write consistency. After a successful PUT, subsequent GETs return the new object immediately. No eventual consistency delays.
Data is replicated across multiple storage nodes in the same datacenter. Designed for high durability, but not geo-redundant by default.
For critical data, combine with off-site backups.Expand volume size without detaching. Filesystem resize handled by guest.
Upgrade storage class on the fly. Migration happens in background.
Point-in-time snapshots. Create new volumes from snapshots instantly.
AES-256 encryption for all data. Keys managed by the platform.
Move volumes between VMs. Data persists independently of instances.
Configurable backup policies. Different retention and replication options.
Manage storage from your terminal
$ scli volume create --size 100 --class standard --name db-data
Creating volume db-data...
ID: vol-x7k2m9p4
Size: 100 GB
Class: standard (10,000 IOPS)
Status: available
$ scli volume attach vol-x7k2m9p4 --vm vm-a1b2c3d4
Attaching volume to vm-a1b2c3d4...
Device: /dev/vdb
Status: attached
$ scli volume resize vol-x7k2m9p4 --size 200
Resizing volume to 200 GB...
Status: resized (run resize2fs in guest)