D-Line 8
$109 /mo
- CPU
- Intel Xeon E-2388G
- Cores
- 8c / 16t @ 3.2 GHz
- Memory
- 64 GB DDR4 ECC
- Storage
- 2 × 960 GB NVMe
- Network
- 1 Gbps unmetered
- IP
- 1 IPv4 + /64
Single-tenant hardware
Whole machines, not slices. Root and IPMI from the first minute, unmetered bandwidth, and a technician in the building if a disk gives up.
Standard builds
In stock means delivery in 2–4 hours. Anything else is a custom build, usually ready within three business days.
$109 /mo
$189 /mo
$429 /mo
$849 /mo
Choosing a build
Most people over-buy cores and under-buy memory. Here is what actually moves the needle.
A database running heavy single-threaded queries wants the highest clock you can get, which is the Ryzen build at 4.5 GHz. Something running hundreds of concurrent jobs — containers, virtual machines, build agents — wants core count instead. Buying 64 slow cores for a workload that only ever uses two of them well is the most common mistake we see.
Memory is the cheapest performance available. A database whose working set fits in RAM is faster than one reading from disk by an order of magnitude, however quick the disk. Offered more cores or more memory for the same money, take the memory.
Every build ships with NVMe, which takes disk off the list of bottlenecks for almost everything. Two drives in RAID1 survive a failure without downtime; four or more in RAID10 give you capacity and redundancy together. Hardware RAID with battery-backed cache earns its price on write-heavy databases and almost nowhere else.
A 1 Gbps unmetered port suits most sites: nothing counted, nothing to overrun. Move to 10 Gbps metered when you are shifting media or backups, where the peak matters more than the monthly total.
Hardware, power and network are ours on every plan. If nobody on your side wants to patch a kernel at midnight, add managed services and the operating system becomes ours too — monitoring, updates, backups and on-call, priced per server.
Included
No line item for the things a server should obviously come with.
Console, power cycling and virtual media over a private management VLAN. Yours, not ticket-gated.
Failed disk, PSU, DIMM or fan replaced within 15 minutes of diagnosis, at any hour.
Always-on scrubbing up to 2 Tbps, with per-service thresholds you can tune with the NOC.
Free layer-2 between your servers, so backups and replication never touch the public port.
Racking, cabling, burn-in and OS install are part of the price.
Debian, Ubuntu, Rocky, AlmaLinux, Windows Server, Proxmox VE or VMware ESXi. Or hand us an ISO.
Details
Typical workloads
PostgreSQL, MySQL and Redis with the working set in RAM and no neighbour competing for the disk queue.
Proxmox or ESXi across 32 or 64 cores, running your own virtual machines with a private VLAN between nodes.
High clock speed, low latency to Israeli players, and scrubbing that stays on through a match rather than dropping the address.
Cores that are always there, so a queued pipeline is a scheduling decision rather than a billing one.
Up to eight NVMe drives per chassis, or a custom build with spinning disks when capacity matters more than speed.
cPanel or Plesk on hardware you control, with root, so you set the limits instead of inheriting ours.
Alternatives
The honest version: dedicated is the right answer less often than we sell it.
| Dedicated | VPS | Colocation | |
|---|---|---|---|
| Whose hardware | Ours, yours alone | Ours, shared chassis | Yours |
| Root access | Yes | Yes | Yes |
| When hardware fails | We replace it within 15 minutes | You are moved, nothing to do | You supply the part, we fit it |
| Scaling up | A new build, 2–4 hours | Resize across one reboot | Buy and ship more |
| Money up front | None | None | The hardware itself |
| Ready in | 2–4 hours | 90 seconds | The same week |
| Worth it when | One tenant needs the whole machine | Load fits in 16 cores | You already own the kit |
FAQ
Yes. Tell us the workload and we will quote a build. Most custom configurations land within three business days.
We manage the hardware, network and power on every plan. Operating system and application management is an add-on — monitoring, patching and on-call, priced per server.
The NOC sees the alert before you do. We contact you, confirm the swap window, and replace it. Data on RAID is rebuilt without downtime.
Ship it and we will rack it as colocation, or we will match the spec on new hardware and give you both machines for a week to migrate.
Yes, on a private management VLAN, with console, power control and virtual media. Rebooting your own server should never require opening a ticket.
Send us an ISO and we will mount it, or mount it yourself over IPMI virtual media and install exactly what you want.
On 1 Gbps ports, yes — we neither count nor shape it. The 10 Gbps plans carry a large allowance instead, because a saturated 10G port is a different economic proposition and pretending otherwise would be dishonest.
The price you sign at holds for the term. We do not raise it on renewal without thirty days notice and the option to leave.
GPU nodes, high-frequency trading builds, 1 PB storage clusters — we quote them all.