VDS Server Hosting

What is a VDS and how is it different from a VPS?

A VDS (Virtual Dedicated Server) is a KVM partition of a physical server; the difference is that the CPU cores and memory allocated to you are never given to another customer. On a VPS, virtual CPUs (vCPUs) run on the host's cores by time-slicing, so when a neighbouring server gets busy, your CPU time can be affected too. A VDS has no such sharing, so you do not get the sudden slowdowns known as the "noisy neighbour" effect.

In practice the difference shows up in consistency rather than average speed: the same query returns in the same time at midnight and at noon, and page generation does not slow down during a sales campaign.

Who is a VDS server the right choice for?

A VDS is the right choice for workloads that use CPU and memory continuously and predictably, where slowness translates directly into lost money or reputation. Typical users are:

• Online stores: shops with heavy catalogue and cart operations on WooCommerce, OpenCart or Magento hosting.
• Database servers: MySQL, MariaDB or PostgreSQL installs whose buffer pool must stay fixed in memory.
• Business applications: Odoo ERP hosting, accounting and inventory software, CRM.
• Files and collaboration: private cloud setups such as Nextcloud hosting.
• Gaming communities: game servers that have no ready-made package or run with lots of mods.
• Agencies: teams that consolidate several client sites with their own panels on one server.

There is a concrete metric that tells you it is time to move from a VPS to a VDS: the steal (st) value in top or vmstat output. It shows how long your virtual CPU was ready to run but waited in line on the host. If it stays above a few percent at busy hours, your application's slowness may come from sharing rather than your code; dedicated cores remove that wait. Likewise, database query times that vary from hour to hour under the same load are a sign of the sharing effect.

If your project is still small, its traffic is uncertain or you are only building a test environment, a VPS server is the more sensible start; you can move to a VDS as you grow.

How do KVM virtualisation and NVMe storage affect performance?

KVM (Kernel-based Virtual Machine) is full virtualisation that gives every virtual server its own kernel and virtual hardware, so your VDS behaves like an independent computer. You can update your own Linux kernel, load kernel modules and run software that needs kernel features, such as Docker and WireGuard, without restrictions. With older container-based methods (OpenVZ, LXC), these operations depend on the host's kernel.

NVMe drives connect to storage directly over PCIe instead of SATA. Where a SATA SSD has a single command queue, NVMe supports thousands of parallel queues; the difference is felt most in the small, random reads and writes (IOPS) that a database performs at the same time. Dedicated cores reduce waiting on the CPU and NVMe reduces waiting on storage; when both bottlenecks are removed together, page generation and query response become more stable.

Which operating systems can you install?

You choose the operating system yourself from the list on the order screen; installation is automatic and access details appear in the customer panel. On the Linux side, long-term-support releases (such as Ubuntu LTS, Debian and AlmaLinux) are the safest choice for most server software; if you want Windows Server, check the listed versions and licence terms on the order screen.

Ask yourself these questions when choosing:

1. Which platform does my software need? ASP.NET Framework, MSSQL or desktop applications need Windows; PHP, Node.js, Python and container stacks need Linux.
2. How long is the support period? LTS releases get security updates for years; short-lived releases soon force an upgrade.
3. Will I install a control panel? cPanel runs on certain RHEL-based distributions and Ubuntu; Plesk runs on both Linux and Windows.

Changing the operating system later requires a reinstall and wipes the disk; take a backup first.

How do you connect to a VDS server?

You connect to a Linux VDS over SSH (Secure Shell) as root, and to a Windows VDS over Remote Desktop (RDP) with the administrator account. The IP address and initial password appear on the service screen of the customer panel once provisioning is complete.

ssh root@SERVER_IP

In the first session, do the following in order: update the system, create a privileged user, switch from password to key-based login, and disable root and password logins.

Root access gives you complete freedom: you install whichever web server (Nginx, Apache, LiteSpeed), database, language version and monitoring tool you like. In return, updates and security hardening are your responsibility; if you would rather hand this work off, you can add the management option when ordering.

How do the network infrastructure and DDoS protection work?

The LIONFIN network connects to multiple carriers with 600 Gbit/s of total upstream capacity, and all server traffic passes through the RouteFence firewall layer. RouteFence filters volumetric DDoS attacks at the network level before they reach your server; bandwidth-flooding attacks such as SYN floods and UDP amplification are stripped out in this layer and clean traffic is forwarded to your VDS.

Network-level protection does not replace a firewall inside the server; the two work together:

• Network layer (RouteFence): handles volumetric and protocol attacks.
• Server layer (set up by you): a UFW firewall that opens only the necessary ports, and Fail2Ban to block brute-force attempts.
• Application layer: a WAF or a caching CDN for websites.

Values such as traffic quota and the number of IPv4 addresses vary by plan and are listed in the current plan table.

Where are the VDS servers hosted, and under what conditions?

Your VDS runs in data center racks cooled with a hot/cold aisle layout; the cold aisle is kept between 18 and 24 °C. This range matches the operating conditions recommended by hardware vendors and prevents heat-related CPU throttling and loss of disk lifespan. See our data center page for details.

The infrastructure is monitored 24/7 by two teams: the NOC (Network Operations Center) watches the network, hardware and availability, and the SOC (Security Operations Center) watches security events. Our target is 99.99% uptime; this is a target, not a contractual guarantee.

The servers are located in Türkiye, which keeps them close to users in Türkiye and the surrounding region; this matters especially for real-time applications, APIs and game servers.

Which VDS plan is enough for you, and how do you size it?

Choose a VDS size with roughly twice your current usage as starting headroom: measure your CPU and memory consumption at the busiest hour and add room for growth. If you already have a server, the output of htop, free -m and iostat, or your control panel's resource graphs, are good starting data.

1. Memory (RAM): add up the database buffer pool, PHP-FPM or application workers and the cache (Redis). If memory runs out, the system swaps to disk and everything slows down.
2. Cores: look at concurrent requests; every active PHP or application worker keeps one core busy.
3. Disk: data + logs + local backup copies + growth. Database files grow over time, and running out of free space is one of the most common causes of outages.
4. Traffic: average page size × monthly page views; on media-heavy sites a CDN reduces traffic.

Concrete core, memory, disk and traffic values are kept current in the plan table on this page. When in doubt, starting one plan smaller and measuring the load is cheaper than leaving a large plan idle; resource upgrades keep your data intact.

How should you set up backups and disaster recovery?

On a VDS you own the data and are responsible for backups; some plans include daily backups (stated on the plan card), but you should also have your own backup plan for critical data. The recommended method is the 3-2-1 rule: three copies of the data, on two different media, one of them off the server.

• Database dumps: nightly backups taken with mysqldump or pg_dump, compressed and copied to another location.
• File backups: incremental backups with restic, borg or rsync, so only changed files are transferred.
• Snapshots: a restore point before major updates; not a replacement for long-term backups.

You only know a backup works by restoring it; run a test restore once a month.

How do you secure a VDS server on day one?

A new VDS becomes a target for automated scanning bots within its first hours online, so security hardening should happen before you install your applications. The following order is a minimum checklist that closes most risks on Linux servers:

1. Update: update all packages with apt update && apt upgrade (Debian/Ubuntu) or dnf upgrade (AlmaLinux) and enable automatic security updates.
2. User and key: create a sudo user, log in with an SSH key and apply PermitRootLogin no and PasswordAuthentication no.
3. Firewall: leave open only the ports you really use (e.g. 22, 80, 443).
4. Brute-force protection: block repeated failed logins with Fail2Ban.
5. Service inventory: list services listening externally with ss -tulpn; internal services such as databases should listen only on 127.0.0.1.
6. Monitoring: set up alerts for disk usage, memory and service status.

On a Windows VDS the equivalents are Windows Update, a strong administrator password, changing the default RDP port or restricting RDP access to specific IPs, and Windows Defender Firewall rules.

Unmanaged VDS or VDS with a management service?

By default a VDS is unmanaged: the hardware, network and virtualisation layer are LIONFIN's responsibility, while the operating system and everything on it are yours. If your team is comfortable on the Linux command line, this is the most flexible and economical model; you make every decision, from versions to configuration.

If nobody will take care of server administration, you can add the management option when ordering; operating system updates and monitoring are then handled by the support team. Answer these three questions to see which model suits you:

• When a security update is released, is there someone to apply it the same day?
• When the disk fills up, a service crashes or a certificate expires, who gets the alert and who acts on it?
• Who regularly tests that backups can actually be restored?

If even one answer is "nobody", a management service usually costs less than the downtime it prevents.

VDS, VPS or dedicated server: what is the growth path?

Server tiers are ordered by how much is shared: VPS (shared vCPU) → VDS (dedicated cores) → dedicated server (the whole machine). Each step adds consistency and isolation and gives up some flexibility and low entry cost.

Signs that you should move from a VDS to a dedicated server: the largest VDS plan is not enough, you need hardware RAID or a specific CPU model, your licence terms depend on physical core count, or you want to run your own virtualisation (Proxmox, Hyper-V). The VDS vs dedicated server comparison covers this decision in detail.

For teams that only use Windows applications over Remote Desktop and do not want to administer a server, an RDP server may be more practical; the difference is explained on the RDP server vs VDS page. AI and rendering work needs a graphics card; in that case look at the GPU server class.