Product reference pages are available in English only.
Storage & Backup

Network File Storage

What Is It?

Network File Storage (NFS) provides shared file storage accessible by multiple VMs using the NFSv4.2 protocol. POSIX-compliant file system with SSD-Standard performance, encryption at rest, and fine-grained access controls. Ideal for shared application data, home directories, and web content.

Quick Facts

Aspect Details
Protocol NFSv4.2
Cluster Size 2 TiB to 42 TiB in 1 TiB increments, all provisioned space usable; can be increased later, never decreased
Performance Class SSD Standard (single tier); service on 2 VMs in Active-Passive mode with a floating IP, storage on two servers in active-active redundancy with RAID
Throughput Up to 300 MB/s per 4 TiB
IOPS 24,000 read IOPS and 18,000 write IOPS per 4 TiB (4 KiB block size), scaling with cluster size
NIC Multi-Queue Activated by default when the cluster has at least two CPU threads; no action needed
File System POSIX-compliant
Encryption At rest (default on SSD-based storage servers)
Security Root-squash (maps client root to anonymous UID, prevents root privileges from clients)
Access Control Per-IP network or host, configurable squash modes per client group
Locations Berlin, Frankfurt, Paris, London, Logrono, Las Vegas, Newark, Lenexa

What You Can Do

Create NFS Clusters

Set up managed NFS clusters with SSD-Standard storage. Fully managed service handles provisioning, redundancy, and maintenance. Clusters support multiple shares with independent access controls.

Configure Shares

Create shares inside a cluster with an optional quota in MiB (0 disables the quota), owner user and group IDs, and access rules. Define which IP networks or hosts can mount each share. Configure squash modes per client group to control user identity mapping.

Mount from Multiple VMs

Multiple Linux VMs, and Kubernetes workloads as shared persistent volumes, can mount the same share simultaneously. Standard NFS mount commands work with NFSv4.2 protocol. Shares accessible across VMs in same data center.

Apply Root-Squash Security

Built-in root-squash maps client root user to anonymous UID, preventing super-user privileges on NFS server. Reduces risk from compromised client machines. Configurable per client group.

Scale Performance

Linear scaling of IOPS and throughput with capacity. Live data movement without interruption. Performance metrics available at file level for monitoring and optimization.

Manage Access

Fine-grained client group access controls. Limit shares to specific IP networks or hosts. Optional squash modes per client group for precise user identity control.

Best For

Scenario Why It Fits
Shared application data Multiple VMs need concurrent read/write access to the same files with POSIX permissions (not for databases or transactional writes)
Home directories Centralized user storage across multiple servers, standard Linux permissions
Web content serving Shared media files across web server pool, NFSv4.2 efficiency, throughput that scales with cluster size (300 MB/s per 4 TiB)
Shared configuration and static assets Multiple VMs share config files, templates, and static assets over POSIX, no per-instance duplication

Key Considerations

Billing & Costs

  • Main billing: By provisioned cluster size, 154 EUR per TB per 30 days (minimum 2 TB, maximum 42 TB); traffic between Network File Storage and your VMs is not charged
  • Performance tier: SSD Standard only (no HDD or SSD Premium option for Network File Storage)
  • Managed service: Redundancy, protocol support, maintenance or encryption, are included in the service.

Limitations

  • NFSv4.2 protocol only (no NFSv3 or SMB/CIFS)
  • Mounting is documented for Linux clients only (VMs and Kubernetes workloads); other client operating systems are not covered
  • Performance scales with capacity (up to 300 MB/s, 24,000 read IOPS and 18,000 write IOPS per 4 TiB)
  • Share access is controlled per client group by IP network or host; inside the share, POSIX permissions apply at file and directory level
  • Same data center access only (no cross-region mounting)
  • Cluster size cannot be decreased after provisioning; it can be increased in 1 TiB steps, so start small and grow.
  • Root-squash security recommended (can be disabled by selecting 'No squash mode', though not recommended for security reasons)
  • Only accessible in private LANs, not from the Internet
  • Not recommended for databases, transactional writes or synchronous write workloads with tight timeouts (write latency is about 20 ms)

Management Options

  • Data Center Designer (DCD)
  • DCD Network File Storage page (create, view, delete clusters; edit name and size)
  • Manage Shares window (create, view, edit, delete shares)
  • Network File Storage API (regional endpoints such as https://nfs.de-fra.ionos.com), SDKs, and Terraform (ionoscloud_nfs_cluster, ionoscloud_nfs_share)
  • Standard NFS mount commands on Linux VMs