4 min read
VDS, VPS and cloud servers compared
What do VDS, VPS and cloud server mean? Compare offers by resource allocation, billing, failure recovery, backups and total cost before you rent a server.
When you search for a server to rent, you find the terms VDS, VPS, virtual server and cloud server. Providers often use them for similar things: a virtual machine that runs on a physical host and receives its own share of resources. The name does not settle the difference. What matters is how resources are allocated, how they are billed and what happens when hardware fails. This guide lists the questions to ask when you compare offers. The Skypium equivalent is cloud servers hosted in Türkiye.
Why the terms overlap
A VPS (virtual private server) is a virtual server created by dividing a physical server with virtualization. Some providers use VDS (virtual dedicated server) to stress that resources are more isolated, but the term has no common technical definition. A cloud server is the same kind of virtual machine running in a resource pool, where you can change the size and pay for usage. Compare the properties below, not the label.
How CPU, memory, disk and network are allocated
| Resource | What to verify |
|---|---|
| vCPU | Number of virtual cores and how physical cores are shared |
| Memory | Allocated memory and how to increase it later |
| Disk | Capacity, storage class, expansion method and separate charges |
| Network | Public IP, private network, bandwidth and traffic charges |
The word “dedicated” alone is not a measurement. Ask the provider for the ratio of vCPUs to physical cores and how servers behave at peak hours. A short test with your own application tells you more than a package label.
Fixed packages or flexible resources
A fixed package gives a set combination of CPU, memory and disk for a monthly amount. With flexible resources you choose the size and pay for the time you use. Hourly billing keeps the cost of short test environments and changing workloads close to actual use. Also ask how resizing works. On Skypium, moving CPU and memory to a larger size stops and restarts the server, so plan it in a maintenance window. Disk capacity is extended separately.
Hardware recovery, high availability and live migration
These three concepts are often confused:
- Hardware failure recovery: if the physical host under your server fails, the virtual machine restarts on a healthy host. Your application does not respond while the server restarts.
- Live migration: when a host needs maintenance, the running virtual machine moves to another host. The server stays on, but there is a short pause at the end of the move.
- High availability: for an application to keep serving when one server is lost, you need more than one server, a copy of the data and traffic routing. A single virtual server does not provide this by itself.
The cloud server page describes Skypium automatic recovery and live migration.
Snapshots and backups are different
A snapshot is a point-in-time copy of a disk. It gives you a quick return point before a critical change. A backup is a copy kept in separate storage under a retention policy. Do not rely on snapshots alone for a disk or storage problem. Review the scheduling and retention options on the backup and DR page. A protection plan is not complete until you have restored a backup in a test.
Which option fits which use case
| Need | Approach |
|---|---|
| Website or small application | Start small, then grow from measured CPU and memory use |
| Test and development | Run hourly resources only when needed; include retained disk and IP items |
| Database | Measure disk latency; compare the maintenance work with a managed database |
| Service sensitive to outages | Design with more than one server and regular, tested backups |
To decide whether a database belongs on your own server or in a managed service, read the guide to managed PostgreSQL versus a virtual server.
Compare more than CPU and memory
Two offers with the same vCPU and memory can have different totals because of storage class, public IPs and backup storage. For each offer, calculate the same workload, the same period and the same protection policy. Use the cloud server total cost guide for the method. Current unit prices and the calculator are on the pricing page. Check the trial scope and test your application before you move production data.