Virtual Dedicated Server Resources Explained
What the resources on a virtual dedicated server really mean: one-to-one vCPUs, reserved RAM, RAID-10 storage, and when a VDS beats a VPS or bare metal.
A virtual dedicated server (VDS) is rented computing that behaves like a machine you own. Its vCPUs are allocated one to one and never oversold, so your speed does not depend on how many servers the host crammed onto the node. It sits between a VPS and bare metal (a full physical server), and the marketing around it is foggy. This post explains each line on a VDS spec sheet in plain terms: what is dedicated, what that changes under load, and when a VDS is the right call.
What is a virtual dedicated server?#
A VDS is a virtual machine whose compute resources are dedicated to you rather than shared with the other tenants on the host. It is still virtualized (you get the reinstalls, rescue mode, and snapshots of a cloud portal), but the cores backing it are not a pool you draw from; they are yours. Think of it as bare metal behavior with virtual machine convenience.
On a VPS, a guaranteed vCore promises you a fair share of a core. On a VDS, every vCPU is allocated one to one against the node's processor capacity, so the node is never sold beyond what it has.
One-to-one vCPUs: the resource that defines a VDS#
Every workload has slow moments, and on shared hardware the usual cause is fighting other tenants for CPU time. When vCPUs are never oversold, the slowest requests look like the average ones: a game server tick, a database commit, or a video encode takes the same time at 3 AM and at peak. That consistency, more than raw speed, is what people are buying when they buy a virtual dedicated server.
It is also the honest version of a promise most of the industry fudges. We publish how our stock control works because dedication only means something if the host enforces it in provisioning, not just in the sales copy.
Memory and storage: reserved, not pooled#
- RAM on a VDS is allocated to your machine in full. None of the quiet tricks that share memory behind your back (overcommit, ballooning), and no slowdown because a neighbor spiked.
- Storage rides RAID-10 arrays: striped for speed, mirrored for survival. A failed drive is our maintenance ticket, not your data loss.
- Transfer allowances on the D line are sized for production use (tens of terabytes monthly); the bandwidth guide explains how allowances and overage work.
The network underneath#
One-to-one vCPUs do not help if the network behind them is thin. VDS nodes live on the same network as everything else we run: 40G spines (the high-capacity switches at the core of our network), transit from multiple independent carriers, and a home at the 1000 Denny Way carrier hotel in Seattle since our March 2026 move, with short undersea-cable paths toward Hawaii, Japan, and Asia-Pacific.
VDS vs VPS vs bare metal: which one do you need?#
| VPS | VDS | Bare metal | |
|---|---|---|---|
| Compute | Guaranteed share of cores | One-to-one vCPUs | The whole machine |
| Consistency under load | High | Highest, never oversold | Highest |
| Reinstall, rescue, snapshots | Minutes, from the portal | Minutes, from the portal | Slower, hardware-dependent |
| Scaling up | Prorated upgrade | Prorated upgrade | New hardware |
| Best for | Most apps and sites | Latency-sensitive and licensed workloads | Special hardware needs |
Rule of thumb: start on a high-clock AMD EPYC VPS unless you know your workload punishes shared compute. Move to a VDS for game server fleets, trading and real-time systems, per-core licensed software, or any production database where tail latency is money. The knowledge base comparison goes deeper on the decision, and what a dedicated CPU VPS really buys you shows how to measure whether you need one.
The current VDS lineup, priced live#
These rows render straight from our pricing data, so they always match the VDS page to the cent, and renewals stay at the price you bought at.
| Plan | vCPUs (one to one) | RAM | RAID-10 storage | Transfer | Monthly |
|---|---|---|---|---|---|
| D40 VDS | 2 | 8 GB | 160 GB | 32 TB | $64.00/mo |
| D50 VDS | 3 | 12 GB | 240 GB | 48 TB | $96.00/mo |
| D60 VDS | 4 | 16 GB | 320 GB | 64 TB | $128.00/mo |
| D70 VDS | 6 | 24 GB | 480 GB | 96 TB | $208.00/mo |
FAQ#
Is a virtual dedicated server the same as a dedicated server?
No. A dedicated server is one physical machine, yours alone. A VDS gives you one-to-one vCPUs inside a virtualized slice of one, so you keep portal-level reinstalls, rescue access, and fast upgrades on a node that is never oversold.
Are VDS vCPUs really dedicated?
On our D line, each vCPU is allocated one to one against node capacity, and that capacity is stock-controlled, so it cannot be resold out from under you. We may rebalance where vCPUs run to keep a node healthy, which is why we do not call them pinned.
Can I run production databases on a VDS?
That is one of the best uses for one. One-to-one vCPUs plus RAID-10 storage keep commit latency flat; pair it with snapshots and off-host backups and you have a serious primary.
How do I move from a VPS to a VDS later?
Order the VDS, migrate, and cancel the VPS; for most stacks it is a rsync plus DNS change, and our engineers will help with the move.