The Open Port Next Door: Why Remote Desktop Is Not a VPN
Photo: Tim Walker from United Kingdom, CC BY 2.0, via Wikimedia Commons
Small accounting firms often describe their remote setup in one sentence: staff log in to the office PC and work. What they mean is that a desktop in the back office is listening on the public internet, sometimes through Remote Desktop Protocol, sometimes through a consumer remote-access tool installed during a busy week and never reviewed. The arrangement feels private because only staff know the address. It is not private. It is exposed.
What an exposure scan actually shows
For this analysis, we reviewed public internet exposure data and ran a controlled scan against a lab network configured like a small firm. We did not scan client networks. The pattern was familiar. An RDP service on port 3389 answered from the public internet. A consumer remote-access agent was installed on a workstation with a weak password and no second factor. A router forwarded traffic to a desktop that had been patched months earlier but never audited.
Internet-wide scanning is not a secret technique. Services such as Shodan and Censys index open ports continuously, and attackers use the same data. An accounting firm does not need to be interesting to be found. It needs only to be reachable. RDP brute-force attempts are automated, relentless, and inexpensive.
The shortcut is understandable, which is why it persists
Remote Desktop is already built into Windows. A consumer remote tool can be installed in minutes. A VPN requires a concentrator, certificates, MFA, and someone to manage it. During busy season, the shortest path wins. The firm gets remote access, the preparer gets to work, and no one documents the decision.
The problem is that the shortcut solves the wrong problem. It gives one person access to one desktop. It does not create a controlled tunnel into the firm's network. It does not enforce device posture. It does not centralize authentication. It does not produce the kind of access log that a client questionnaire or cyber-insurance underwriter will accept.
Why it fails the reviews that matter
Client security questionnaires increasingly ask how remote users authenticate, whether access is logged, and whether the firm uses MFA at the network boundary. A direct RDP exposure answers those questions badly. Even if the firm uses a strong password, the service is still reachable by anyone on the internet. A consumer remote tool may encrypt the session, but it often bypasses the firm's VPN, identity provider, and logging stack.
Cyber-insurance carriers have also tightened their questions. They ask about remote access, MFA, and administrative controls. A firm that cannot describe a tunneled, MFA-protected entry point may face higher premiums, exclusions, or a declined renewal. The issue is not the wording of the questionnaire. The issue is that the underlying architecture is difficult to defend.
What a proper remote-access setup looks like
The replacement does not need to be enterprise-grade. It needs to be deliberate.
- Terminate remote access at a VPN concentrator or a zero-trust access gateway, not at a desktop.
- Require MFA at the tunnel, not only at the tax application.
- Issue named accounts. Shared logins defeat accountability.
- Enforce device posture where possible: managed laptop, current operating system, disk encryption, and endpoint protection.
- Limit access to the systems each role needs. A seasonal preparer does not need the entire network.
- Log authentication events and retain them long enough to support incident review and client inquiries.
A small firm can run this model with a business-grade router, a subscription VPN service designed for teams, or a managed service provider. The point is not the brand. The point is that the entry point is controlled, authenticated, and observable.
A migration sequence before January 2027
Start by inventorying every remote-access path: RDP forwards, consumer remote tools, old VPN accounts, and any router rules that point to a workstation. Disable public RDP first. Then stand up the replacement tunnel and test it with a small group. Move users in batches, not all at once. Rotate credentials that were used under the old model. Finally, document the new architecture in plain language for questionnaires and insurance renewals.
The firm does not need a perfect security program. It needs to stop treating a desktop on the internet as a private door. Remote Desktop is a protocol, not a VPN. The distinction matters to clients, carriers, and anyone who understands what an open port means.