An RDP VDS is a virtual server running Windows that you reach over the Remote Desktop Protocol. You open a window on your own computer, see the desktop of a machine sitting in a data center, and use it as if it were in front of you. This guide explains what an RDP VDS actually is, who benefits from one, how to size it for your workload and, most importantly, how to protect an internet-facing remote desktop from attack, one step at a time.
Technically there is no separate product category. A VDS is a virtual server created by dividing a physical machine with KVM virtualization, with its own CPU cores, RAM, disk and IP address. Once Windows is installed on it, the natural way to manage it is Remote Desktop, so people simply call it an RDP VDS or an RDP server. You reach a Linux VDS through SSH on the command line; you reach a Windows one through RDP with a full graphical desktop. The difference lies in the operating system and the way you connect, not in the hardware.
RDP is built into Windows and listens on port 3389 by default. You do not need extra software on the client side: Windows ships with Remote Desktop Connection (mstsc), Microsoft publishes official apps for macOS, Android and iOS, and Linux users typically rely on clients such as Remmina.
What separates an RDP VDS from remoting into an office PC is continuity. The server runs around the clock on redundant power and network infrastructure, and when you close the connection your session and your running programs stay exactly where you left them.
The range of uses is wider than most people expect. These are the scenarios we see most often:
An RDP VDS is a general-purpose Windows server. Its use must comply with the service agreement and applicable law; activities such as bulk unsolicited email, scanning other people's systems or unauthorized access attempts lead to termination of the service.
The Windows graphical interface needs more memory than a headless Linux system. The operating system alone takes a few gigabytes of RAM, and your applications come on top of that. Treat the table below as a starting point rather than a precise rule. Because a KVM-based VDS can be upgraded later, the sensible approach is to start small and watch actual resource usage.
| Workload | vCPU | RAM | Note |
|---|---|---|---|
| One user, one light application | 2 | 4 GB | Background bot, simple automation |
| One user, browser and office work | 2-4 | 8 GB | Browser tabs consume memory quickly |
| Accounting software, 3-5 users | 4 | 8-16 GB | Allow 1-2 GB per user |
| Web application on IIS and SQL Server | 4-8 | 16 GB or more | Set a maximum memory limit for SQL Server |
| 10 or more concurrent users | 8 or more | 32 GB or more | A dedicated server is worth considering here |
For storage, set aside at least 50-60 GB for the operating system and its updates; SSD or NVMe storage noticeably shortens sign-in and application start-up times. For many concurrent users or heavy databases, dedicated server options may be a better fit. If you are undecided, see our comparison of virtual and dedicated servers.
SistemDC installs Windows Server 2016, 2019, 2022 and 2025 as well as Windows 11 from ready-made images. Work through these questions in order:
For most new deployments Windows Server 2022 or 2025 is the right choice. For a closer look, read the Windows Server 2022 VDS guide and the Windows Server 2025 VDS guide.
Deploy your Windows VDS in minutes
Hosted in Istanbul on KVM virtualization with a 1 Gbps port; Windows Server 2016-2025 and Windows 11 images ready to install.
Browse Windows VDS plansThis is the most commonly misunderstood point. Windows Server allows two simultaneous remote sessions for administration. Two people managing the server can therefore connect without any extra setup.
If three or more people need to work at the same time, or the server is going to act as a shared desktop for a team, you must install the Remote Desktop Services (RDS) role and obtain an RDS client access license (CAL) for each user or device. Licensing rules are set by Microsoft and can change over time; check Microsoft's RDS licensing documentation for the current terms. Settling your user count before you pick a plan makes both capacity and licensing easier to get right.
mstsc and press Enter.IP:port.Closing the window does not end the session; your programs keep running. To list open sessions on the server, run:
quser
An open port 3389 is one of the targets automated scanners probe most. It is normal for a newly installed server to log hundreds of failed sign-in attempts within a few hours. The goal of those attempts is to guess a weak password and, more often than not, to plant ransomware afterwards. Everything below takes less than half an hour.
This is the single most effective measure. If your office has a static IP, narrow the firewall rule so that only that address is allowed. The example uses a documentation address; replace it with your own:
Set-NetFirewallRule -DisplayGroup "Remote Desktop" -RemoteAddress 203.0.113.10
If you do not have a static IP, the VPN approach in step 10 achieves the same result.
With NLA on, the server does not even create a sign-in screen until the user has authenticated, which reduces both resource use and attack surface. It is enabled by default on current releases, but verify it:
Set-ItemProperty -Path "HKLM:\System\CurrentControlSet\Control\Terminal Server\WinStations\RDP-Tcp" -Name UserAuthentication -Value 1
Attack tools try the name "Administrator" first. Create a new administrator account with a different name or rename the built-in one, and set a password of at least 14 characters that is not a dictionary phrase and is not used anywhere else.
Rename-LocalUser -Name "Administrator" -NewName "server-admin"
For daily work, create separate accounts without administrator rights and add only the people who really need access to the Remote Desktop Users group.
Locking an account for a while after repeated wrong passwords makes brute-force attempts pointless. The following command locks an account for 30 minutes after 5 failed attempts:
net accounts /lockoutthreshold:5 /lockoutduration:30 /lockoutwindow:30
A custom port is not a security control, but it cuts out most automated scanning and keeps your logs readable. Order matters: add the firewall rule for the new port first, then change the port.
New-NetFirewallRule -DisplayName "RDP-Custom-TCP" -Direction Inbound -Protocol TCP -LocalPort 33890 -Action Allow
New-NetFirewallRule -DisplayName "RDP-Custom-UDP" -Direction Inbound -Protocol UDP -LocalPort 33890 -Action Allow
Set-ItemProperty -Path "HKLM:\System\CurrentControlSet\Control\Terminal Server\WinStations\RDP-Tcp" -Name PortNumber -Value 33890
Restart-Service TermService -Force
Your connection drops when the service restarts; reconnect on the new port. If something goes wrong, you can reach the server screen through the console in the Client Area and fix the setting. Once you have confirmed the new port works, disable the old 3389 rules.
Remote Desktop components have had serious vulnerabilities in the past that required no authentication at all. Keep Windows Update enabled, install the monthly security updates and schedule the restart outside working hours.
Disconnected sessions left open for days consume memory and add risk. In gpedit.msc, go to Computer Configuration > Administrative Templates > Windows Components > Remote Desktop Services > Remote Desktop Session Host > Session Time Limits and set limits for idle and disconnected sessions. If a program has to run continuously, run it as a service or a scheduled task rather than leaving it tied to a session.
By default your local drives and clipboard can be carried into the session. Malware that reaches the server can then reach your mapped local drives as well. If you do not need it, turn off drive redirection under Device and Resource Redirection in the same policy tree.
Failed attempts are written to the Security log with event ID 4625 and successful sign-ins with 4624. For a quick look at the most recent failures:
Get-WinEvent -FilterHashtable @{LogName='Security'; Id=4625} -MaxEvents 20 | Format-Table TimeCreated, Message -Wrap
If you see a successful sign-in from an address you do not recognize, change the passwords immediately and investigate the server.
The most robust design is one where the RDP port is not exposed to the internet at all. You install a VPN such as WireGuard on the server or on a small separate VDS and allow RDP only from the VPN network. For teams without a static IP, working in the field or from home, this is the most practical answer. Larger multi-user environments should also consider Remote Desktop Gateway and multi-factor authentication.
Despite every precaution, an account can be compromised, an update can break the system, or a user can delete the wrong folder. That is why backups belong on the security list:
SistemDC offers an optional, paid Backup Service add-on for VDS plans, which can be added at order time or later.
SistemDC VDS plans run in an Istanbul data center built to Tier 3 standards, on KVM virtualization with a 1 Gbps port; restart, reinstall and console operations are handled through the Client Area. For current plans and pricing, see the VDS hosting page.
A Windows VDS ready for Remote Desktop
Pick your version, have the server provisioned and connect over RDP. Our support team accepts tickets around the clock.
See Windows VDS plansIs an RDP VDS the same thing as a Windows VDS?
For practical purposes, yes. RDP VDS is the everyday name for a Windows virtual server that you reach over the Remote Desktop Protocol. It is not a separate product type; every VDS running Windows is accessed through RDP.
How many people can connect to an RDP VDS at once?
Windows Server allows two simultaneous remote sessions for administration. To let more people work at the same time you need to install the Remote Desktop Services role and obtain an RDS client access license per user or per device.
Does changing the RDP port make the server secure?
Not on its own. A custom port removes most automated scanning and keeps your logs readable, but a targeted attacker can still find it. Real protection comes from combining IP restrictions, NLA, strong passwords and an account lockout policy.
How much RAM does an RDP VDS need?
For one user and a light application, 4 GB is a reasonable starting point. Browser-heavy work, accounting software or a few concurrent users call for 8 GB or more, and servers that also run a database should plan for 16 GB or more.
Do my programs stop if the RDP connection drops?
No. When the connection drops or you close the window, the session stays open in a disconnected state and your programs keep running. They only stop if you choose to sign out of the session.
How do I reach the server if I lock myself out of RDP?
The console in the Client Area gives you access to the server screen independently of its network settings. If you make a mistake in a firewall rule or port setting, you can log in through the console and correct it.