Remote Access Done Right: Replacing Exposed RDP With a Secure VPN
Exposed RDP is one of the easiest ways into a small business network. Here is how to find it and replace it with secure remote access that works.
It's Monday morning. Your production planner can't log in, and the reason turns out to be a Windows account that got locked out over the weekend by thousands of failed password attempts from IP addresses in other countries. Someone, years ago, opened port 3389 on the firewall so the owner could reach his office PC from home. Nobody closed it. That is the most common way I see small companies get secure remote access wrong: it works, so nobody looks at it again.
I've been a network engineer for 15 years, and exposed Remote Desktop (RDP) is still one of the first things my partner Mukesh and I look for when we assess a new customer's network. It's also one of the easiest fixes to get right, as long as you plan it instead of ripping it out on a Friday afternoon.
This guide walks through why exposed RDP is so dangerous, how to find out if you have it, and how to replace it with a VPN (or something better) without breaking the way your people work.
In this guide
- Why exposed RDP is a problem
- How to check if RDP is open to the internet
- Your options for secure remote access
- How to replace exposed RDP with a secure VPN, step by step
- Hardening the VPN itself
- Remote access for vendors and machine builders
- Frequently asked questions
Why exposed RDP is a problem
RDP is a perfectly good tool inside your network. The problem is putting it directly on the internet, usually by forwarding TCP port 3389 (or a "hidden" port like 33890) from the firewall to an internal PC or server.
The moment that port is reachable, automated scanners find it, often within hours. Then the password guessing starts and never really stops. If one account has a weak or reused password, an attacker is sitting at a Windows desktop inside your network.
Incident response firms have reported for years that RDP abuse shows up in a large share of ransomware cases, both as the way in and as the way attackers move between machines once inside. Exposed RDP has also had serious vulnerabilities of its own over the years, BlueKeep being the best known, that could be attacked without any password at all.
Why "we changed the port" doesn't help
Moving RDP from 3389 to some other number cuts down the noise in your logs. It doesn't stop a determined scanner, which checks every port. Treat a non-standard port as a cosmetic change, not a security control.
What it costs you beyond the breach
Cyber insurance applications now commonly ask whether RDP is exposed to the internet and whether remote access requires multi-factor authentication. A "yes, it's open" answer can mean higher premiums, exclusions, or a declined application. I covered how to handle those questions in our guide to cyber insurance security questionnaires.
How to check if RDP is open to the internet
You don't need special tools to start. Here's the checklist I use:
- Read your firewall's port forwarding (NAT) rules. Look for anything forwarding to TCP 3389, or any rule pointing at an internal workstation or server you don't recognize.
- Check every internet connection. Many plants have a second ISP line, a cellular backup, or an old DSL modem that a vendor installed. Each one is a separate front door.
- Scan from outside. From a device that is not on your network, run a port scan against your public IP addresses (only addresses you own and are authorized to test). Free tools like Nmap work fine.
- Look at public search engines. Services such as Shodan index internet-facing devices. Searching your public IP shows what the rest of the world already sees.
- Ask your people. Remote-control apps installed without IT's knowledge are the modern version of the open port. Ask who connects from home and how.
If you find exposed RDP, don't panic and don't just yank the rule. Someone depends on it. Find out who, then move on to the plan below.
Signs it's already being attacked
Check the Security event log on the exposed machine for large numbers of failed logons (Event ID 4625) from outside IP addresses. Frequent account lockouts with no obvious cause are another clue. If you see successful logons (Event ID 4624, logon type 10) from places you can't explain, treat it as an incident, not a cleanup project.
Your options for secure remote access
There's no single right answer. Here is how the common options compare for a company under 500 people.
Option 1: A remote access VPN
The user runs a VPN client on their laptop, authenticates (with MFA), and gets an encrypted tunnel into the network. Then they use RDP over that tunnel, the same way they would in the office. Most business firewalls already include a VPN feature, so this is usually the fastest and cheapest fix.
Good fit when: you have a handful to a few dozen remote users, a modern firewall, and staff who need access to internal file shares or apps.
Option 2: Remote Desktop Gateway
Microsoft's RD Gateway wraps RDP inside HTTPS and lets you put access policies and MFA in front of it. It's a reasonable choice for companies that run a lot of Windows servers, but it's one more internet-facing server to patch and monitor.
Option 3: Zero trust network access (ZTNA)
ZTNA services connect a verified user to one specific application or machine, not the whole network. Nothing on your side listens for inbound connections from the internet. Pricing is typically per user per month, and it's often a strong option for companies with many remote users or several sites.
Option 4: Managed remote support tools
Tools that let IT staff or an MSP take control of a PC are fine for support, but they are not a substitute for a remote access strategy for employees. Lock them down with MFA and limit who has accounts.
My usual recommendation: for most small manufacturers I talk to, a firewall-based VPN with MFA is the right first step. It closes the hole quickly. ZTNA is often the better long-term destination once the urgent risk is gone.
How to replace exposed RDP with a secure VPN, step by step
Here's the sequence we follow. It's built so nobody loses access in the middle of a shift.
Step 1: Make a list of who uses remote access, and for what
Write down each person, what they connect to (their desk PC, a server, the ERP system), and when. You'll almost always find a couple of accounts nobody can explain. Those get disabled.
Step 2: Confirm your firewall can do it
Check that the firewall is still supported by the vendor, is on current firmware, and is licensed for the number of VPN users you need. If it's past end of support, fix that first. Our post on signs your firewall is outdated covers what to look for.
Step 3: Set up MFA before anyone connects
The VPN should require a second factor for every user, every time. Tie it to your existing identity system (for many companies, that's Microsoft Entra ID) so that when someone leaves, disabling one account cuts off everything.
Step 4: Limit what VPN users can reach
A VPN user should not land on a flat network where every PC, printer and PLC is reachable. Put VPN clients in their own address range and write firewall rules that allow only what each group needs. If your network is flat today, network segmentation is the next project.
Step 5: Pilot with two or three users
Pick one person in the office and one who travels. Have them work over the VPN for a few days. Fix the small problems (DNS names that don't resolve, mapped drives, printing) before everyone else moves.
Step 6: Roll out, then close the port
Move the rest of the users over, confirm each one works, and then remove the port forwarding rule. Scan from outside again to confirm 3389 is closed on every public IP. Put a date on the calendar to re-check in 30 days, because old rules have a way of coming back.
Step 7: Reset passwords on the exposed machines
Assume any account that could log on through the exposed port has been guessed at. Reset those passwords, and if you see anything suspicious in the logs, have the machines checked properly before calling it done.
Hardening the VPN itself
A VPN is not a set-and-forget appliance. Over the last few years, VPN and firewall products from several major vendors have had serious, actively exploited vulnerabilities, and CISA has repeatedly added them to its Known Exploited Vulnerabilities catalog. Your VPN is now your front door, so treat it that way.
- Patch fast. Subscribe to your firewall vendor's security advisories. Critical VPN patches should go in within days, not at the next quarterly maintenance window.
- Turn off what you don't use. If you only use the client VPN, disable unused web portals and management interfaces on the internet-facing side.
- Never expose management to the internet. Firewall admin pages should only be reachable from inside the network or a dedicated management connection.
- Log and alert. Send VPN logs somewhere someone actually reviews. Logins at 3 a.m. from a country where you have no employees deserve a phone call.
- Protect the endpoints. A laptop on the VPN is part of your network. It needs current patches and good endpoint protection, ideally EDR rather than basic antivirus alone.
- Review accounts quarterly. Remove people who left, vendors who finished their project, and anyone who hasn't connected in 90 days.
A note for defense suppliers
If you handle Controlled Unclassified Information, NIST SP 800-171 includes requirements to control and monitor remote access and to use multi-factor authentication. The VPN design above supports those goals, but confirm the details for your situation with your assessor.
Remote access for vendors and machine builders
On the plant floor, the riskiest remote access often isn't your employees. It's the equipment vendor who installed a cellular modem or remote-control app on an HMI so they can troubleshoot without driving in.
You probably can't eliminate vendor access, and you shouldn't try. Production needs it. You can put rules around it:
- Find every vendor connection. Walk the floor and look for unfamiliar modems, routers and remote-control software on HMIs and engineering workstations.
- Route it through one controlled path. Vendors connect through your VPN or a dedicated access portal, with their own named accounts and MFA.
- Turn it on only when needed. Vendor accounts stay disabled until someone on your team enables them for a scheduled session.
- Restrict them to their own equipment. The robot vendor's account reaches the robot cell, nothing else.
- Keep a log. Record who connected, when, and what they worked on.
This matters even more when those machines run older Windows versions that can't be patched. I wrote about that situation in legacy Windows machines on the plant floor.
Frequently asked questions
Is it safe to use RDP over the internet if I have a strong password?
A strong password helps, but it doesn't protect you from RDP vulnerabilities that don't need a password, and it doesn't stop a password that was reused and leaked elsewhere. Put RDP behind a VPN, gateway or ZTNA service with MFA instead of exposing it directly.
What's the difference between a VPN and RDP?
RDP is the tool that lets you see and control a remote Windows desktop. A VPN is the secure tunnel that puts your laptop on the company network first. The safe pattern is to connect the VPN, then use RDP through it.
How much does secure remote access cost for a small business?
If your firewall is current and includes VPN licensing, the main costs are setup time and an MFA solution, which many companies already have through Microsoft 365. ZTNA services are commonly priced per user per month. Replacing an end-of-life firewall is usually the largest expense, and the price varies a lot with your bandwidth and feature needs.
Do I still need a VPN if we use Microsoft 365?
Not for email, Teams or SharePoint, since those are already cloud services. You still need secure remote access for anything that lives on-site, like file servers, ERP systems, engineering workstations or plant equipment.
Is a VPN enough, or do I need zero trust?
For many small companies, a well-run VPN with MFA and tight access rules is enough to close the immediate risk. Zero trust tools add finer control and remove the internet-facing VPN entirely, which is worth considering as you grow or add sites.
Not sure what's exposed? Let's find out
Most of the companies we talk to don't know exactly what's open on their firewall until someone checks. That's normal, and it's fixable.
Our free cybersecurity assessment includes a look at your remote access setup: open ports, VPN configuration, MFA, and vendor connections. You'll get a plain-English list of what to fix first. If you'd like us to take a look, reach out here and we'll set up a time.
Related reading
- Signs Your Firewall Is Outdated (and What to Replace It With)
- Network Segmentation for Small Businesses: A Practical Guide
- What to Do in the First Hour After Someone Clicks a Phishing Link




