🏗️

Cold Data · Archive · Media · JBOD · 6 Locations

Bulk Storage
Bare Metal Servers

High-density bare metal storage for cold data lakes, long-term archives, media repositories, and object storage backends — up to 192 TB raw in a single server with 10 Gbps unmetered uplink and full IPMI access. Available across 6 datacenters in Frankfurt, Strasbourg, New York, Miami, and Singapore. From €199/mo.

192

Max TB Raw

€199

From /mo

10G

Uplink

Locations

Zen 4

Architecture

Browse Bulk Plans Speak With an Engineer

RAID 6 or JBOD / ZFS

10 Gbps Unmetered

Up to 25 Gbps Uplink

48h Deploy SLA

🏗️

Bulk Storage Server

12× 16 TB SATA HDD · RAID 6 / JBOD · 10 Gbps

Drive Config 12× 16 TB Enterprise SATA HDD Core Range 32– to 64 cores (1P) RAID RAID 6 (default) · JBOD for ZFS Max RAM (1P) 1152 GB per socket Storage PCIe Gen 4 NVMe · Up to 15.4 TB Network Options 1 Gbps or 25 Gbps Unmetered

Why Bulk Storage Bare Metal

When your dataset is measured in tens or hundreds of terabytes, the economics of cloud cold storage collapse quickly. AWS Glacier charges ~$4/TB/month in storage alone — but adds $20–$90/TB in retrieval fees that make large restores prohibitively expensive. A bulk storage server at €199/mo delivers 192 TB raw (~160 TB usable in RAID 6) at €1.24/TB/month with no retrieval fees and instant access at 10 Gbps. You control the RAID topology, filesystem, and access model — deploy MinIO for S3-compatible object storage, use ZFS with JBOD for maximum flexibility, or run a simple XFS mount for sequential media workloads. IPMI gives you full out-of-band access without opening support tickets.

12× 16 TB Enterprise SATA HDD

192 TB raw · ~160 TB usable (RAID 6)

DDR5-4800 ECC Registered

12 channels · Max 1152 GB

10 Gbps Unmetered

No egress charges · ~3.24 PB/mo max

Cost Comparison

Bulk Storage vs Cloud Cold Storage

Specification	CoreNetHub Bulk Storage	AWS S3 Standard	AWS Glacier
Cost at 100 TB stored	€199/mo flat	~$2,300/mo + egress	~$400/mo + retrieval
Retrieval speed	Instant · 10 Gbps	Fast (egress fees apply)	3–12 hours
Egress / retrieval fees	None	$90/TB out	$20–$90/TB out
Max capacity	192 TB raw per server	Unlimited (at cost)	Unlimited (at cost)
Software stack	Any (MinIO, ZFS, XFS…)	S3 API only	S3 API only
Cost/TB/month (100 TB)	€1.99/TB	$23+/TB	$4+/TB (+ retrieval)

Use Case Fit

Built for High-Volume Data Workloads

Bulk storage bare metal makes sense when your dataset is large enough that per-TB cloud pricing becomes the dominant cost — and when instant, fee-free retrieval matters.

Cold Data Lakes & Archive

Long-term retention of logs, analytics data, compliance records, and historical datasets that are rarely read but must remain accessible. At €1.99/TB/month (RAID 6 usable), bulk storage is 10× cheaper than S3 Standard and delivers instant retrieval at 10 Gbps — compared to AWS Glacier's 3–12 hour restore delay and $20–$90/TB retrieval fee.

€1.99/TB/mo · instant access

S3-Compatible Object Storage

Deploy MinIO on a bulk storage server to create an S3-compatible object storage endpoint — fully API-compatible with AWS S3, supporting bucket policies, versioning, and lifecycle rules. A single 192 TB server running MinIO gives you a private S3 backend at a flat monthly rate, with no per-request fees and no ingress charges for data written by your applications.

MinIO · S3-compatible API

Video Surveillance & Media Repositories

CCTV footage, broadcast media, raw video production files, and rendered assets all grow without bound. A bulk storage server with 160 TB usable holds months of high-resolution surveillance footage or a complete broadcast media library. Sequential read throughput from 12 HDDs saturates a 10 Gbps uplink for streaming or CDN origin workflows.

160 TB usable · 10 Gbps

Ceph & Distributed Storage Nodes

Running a Ceph cluster or GlusterFS deployment? Bulk storage servers make ideal OSD (Object Storage Daemon) nodes — high drive count, dedicated CPU (16-core EPYC 7302P), and 10 Gbps uplink for intra-cluster replication. JBOD mode gives Ceph direct access to individual drives for its own failure domain management.

JBOD mode · 10 Gbps cluster

Configuration Guide

RAID 6 or JBOD — Which Is Right for You?

Two deployment modes for the same hardware — choose based on whether you want hardware-managed redundancy or software-defined storage.

RAID 6 — Simplest to Operate

12× 16 TB HDD ~160 TB usable Dual-drive fault tolerance

RAID 6 is the recommended default for most workloads. The hardware RAID controller manages all drive failures transparently — when a drive fails, you get an alert via IPMI and the array continues operating. Usable capacity is ~160 TB (192 TB raw minus 2-drive parity overhead). Best for: media repositories, log archives, cold data lakes, backup targets, and any workload where operational simplicity matters more than maximum raw capacity.

View RAID 6 plans

JBOD — Maximum Flexibility

12× 16 TB HDD 192 TB raw Software-defined RAID

JBOD mode presents all 12 drives as individual devices to the OS, giving your software full control over the storage layer. This is the right choice for ZFS (OpenZFS or TrueNAS) — which implements its own redundancy and integrity checks at the filesystem layer — and for Ceph OSD deployments, MinIO erasure coding, and any application that manages its own data placement. You get 192 TB of raw drive capacity to partition however your software requires.

View JBOD plans

Need Less Capacity?

Backup: 32–48 TB raw NAS: up to 96 TB From €89/mo

If 192 TB is more than you need, our Backup and NAS server lines offer lower-capacity configs starting from €89/mo. Backup servers (32–48 TB raw, RAID 6, 1 Gbps) are the most cost-efficient offsite target for Veeam and rsync workloads. NAS servers (up to 96 TB with SSD cache, 10 Gbps) add a performance cache tier for mixed read/write workloads and shared file serving.

Compare all storage plans

Bulk Storage FAQ

Common Questions About Bulk Storage Servers

RAID 6 uses the equivalent of two drives' worth of capacity for parity data distributed across all drives. With 12× 16 TB drives, you have 192 TB raw. RAID 6 reserves 2 drive equivalents for parity: 192 TB − (2 × 16 TB) = 160 TB usable. This gives you dual-drive fault tolerance — the array survives any two simultaneous drive failures without data loss. If you choose JBOD mode, all 192 TB is presented raw to the OS for software-managed storage (ZFS, Ceph, MinIO).

Choose RAID 6 if you want the simplest setup and don't need to manage storage at the OS level — install your OS, mount the array, and start using 160 TB. Choose JBOD if you're deploying ZFS (TrueNAS, OpenZFS), Ceph, or MinIO erasure coding — these tools implement their own redundancy and integrity checks that work better with direct drive access than on top of a hardware RAID layer. ZFS in particular benefits from JBOD: it can detect silent data corruption per-drive, manage scrubbing individually, and replace failed drives without RAID controller interaction.

12× enterprise SATA HDDs in RAID 6 deliver approximately 1.2–1.8 GB/s sequential read (aggregate) depending on access pattern and RAID controller cache. Individual drive sequential read is ~250 MB/s for modern 16 TB CMR HDDs. The 10 Gbps network uplink (~1.25 GB/s theoretical max) is roughly matched to the array's sequential read throughput — meaning you can saturate the uplink with sequential reads. Random IOPS are much lower: expect 800–1,200 IOPS for a 12-drive RAID 6 array. Bulk storage servers are optimised for sequential, not random, access.

Yes. MinIO is a popular choice for bulk storage servers — it exposes an S3-compatible API on top of the local filesystem, letting your applications use standard S3 SDK calls while data stays on-premise (or in your chosen datacenter). For MinIO, we recommend RAID 6 mode (simpler to operate) or erasure coding across drives in JBOD mode for more granular control. MinIO on a 160 TB RAID 6 array gives you a private S3 bucket with ~1.2 GB/s aggregate throughput and no per-request or per-GB charges.

All bulk storage configs ship with an AMD EPYC 7302P (16 cores / 32 threads, 3.00 GHz base) and 128 GB DDR4 ECC RAM. The 16-core EPYC 7302P is well-matched to storage workloads — it handles ZFS scrubbing, MinIO erasure coding, Ceph OSD processes, and concurrent rsync/scp sessions without CPU contention. 128 GB ECC RAM is sufficient for ZFS ARC caching (ZFS can use up to 50% of RAM as read cache by default) while leaving ample headroom for application processes running alongside the storage layer.

The 25G suffix means this configuration ships with a 25 Gbps network uplink instead of the standard 1 Gbps. Bandwidth remains unmetered in both cases. The 25G config also includes the maximum memory (1152 GB DDR5 ECC) and 4× 3.84 TB NVMe (15.4 TB raw). It’s designed for workloads where both internal throughput and external bandwidth are bottlenecks simultaneously: LLM model serving with large batch sizes, real-time streaming analytics, or CDN/object storage backends that need to saturate network bandwidth while keeping a large working set in RAM.

Bulk Storage
Bare Metal Servers

Bulk Storage vs Cloud Cold Storage

Bulk Storage Plans. 6 Global Locations.

Built for High-Volume Data Workloads

RAID 6 or JBOD — Which Is Right for You?

Common Questions About Bulk Storage Servers

Need Less Capacity or Shared File Access?

Bulk StorageBare Metal Servers

Bulk Storage vs Cloud Cold Storage

Bulk Storage Plans. 6 Global Locations.

Built for High-Volume Data Workloads

RAID 6 or JBOD — Which Is Right for You?

Common Questions About Bulk Storage Servers

Need Less Capacity or Shared File Access?

Bulk Storage
Bare Metal Servers