GUIDE · VULNERABILITY AND TESTING

Unsupported (end-of-life) systems: how to manage them when they cannot be updated

By the end of this guide you will know how to handle a system the vendor no longer updates and that you cannot replace straight away. You will see how to classify it, which compensating measures to apply, how to verify them, and how to record the exception with an exit plan that has owners and a date.

Updated october 2026
7 min read
Altovar Technical Editorial Team

1. Why an unsupported system is a problem

An unsupported system is a product that its vendor no longer updates. It receives no security fixes, so any flaw discovered after support ends stays open unless someone intervenes with alternative measures. This is the typical case with old servers, industrial equipment, abandoned management software and devices running an operating system that is no longer maintained. Even a system that seems secondary can become the entry point to the main network.

The problem is not only technical. These systems are often tied to processes that cannot stop, to tools that have no ready replacement, or to suppliers who no longer respond. Not updating is therefore not a neutral choice: it must be managed as a risk, with owners, measures and a date. If the system processes personal data, managing the risk also becomes a data protection obligation.

2. What you need before acting

Three things are needed before any intervention. The first is an inventory stating which systems are unsupported, with the end-of-support date given by the vendor and the owner of each. The second is a map of dependencies: which applications, users and networks communicate with that system, and what happens if it stops. The third is knowing what data it holds or processes.

Without this information, decisions become arbitrary. An isolated system with no data can be managed differently from one that feeds your invoicing. Ask the vendors of the installed software what alternatives they offer and on what terms, because the answer often decides the most practical route and the time needed to get out. Record the answers in writing, as they will form the basis of the decision and of the budget request.

3. The steps to follow

The sequence below is the basis for managing every system in this condition. You can apply it even when replacement is already scheduled, because the interim measures remain valid until the exit date. Do not wait for the replacement plan to be approved before securing the system: isolation and access control can be activated straight away, often with few changes.

  1. Identify the system and confirm it in the inventory, with owner, function and end-of-support date.
  2. Classify it by business impact and type of data, so you know what level of intervention it merits.
  3. Isolate the system from the general network, limiting connections to those strictly necessary.
  4. Apply compensating controls: named accounts, reduced privileges, event logging and monitoring.
  5. Verify that a backup of the system exists, is separate and has been tested with a restore.
  6. Write a replacement plan with an owner, an indicative budget and an exit date, and record it as an approved exception.
  7. Review the exception on the planned date, and update the classification if the risk has changed.

4. Common mistakes to avoid

The most frequent mistakes arise from haste or from wanting to close the problem quickly. In many cases the system stays in production for years without an explicit decision, and the risk grows unnoticed. Writing one line of decision for each exception is the simplest way to avoid this, and it costs far less than correcting a mistake after an incident.

  • Leaving the system on the general network because it has always been that way.
  • Relying on an antivirus as the only protection, without reducing connections.
  • Running backups without ever testing a restore.
  • Accepting the risk verbally, without a name, reason and review date.
  • Waiting for a failure before deciding on replacement, when purchase lead times have already run out.

5. How to verify that the measures work

A compensating measure must be verified, not assumed. Check that the isolation is effective: from a workstation outside the segment, the system must not be reachable on ports it does not need. Check that access is by named accounts, that unused accounts are disabled, and that access logs reach a place where someone reads them. If logs end up somewhere nobody checks, the measure exists only on paper.

Also run a backup restore test on a separate environment, and record how long it takes and which steps are missing. Finally, review the list of who can access the system, because permissions tend to accumulate. If any of these checks fails, the residual risk is higher than recorded and must be updated in the register. A risk that changes without being reflected in the register becomes invisible again to those making decisions.

6. What to do next: replacement and contracts

The temporary measure must always be accompanied by an exit plan. The plan describes the target system, the migration activities, the people involved and the conditions for closing the exception. If replacement requires investment, present it as risk reduction, using the elements management understands: possible production shutdown, data involved, obligations to customers. This kind of presentation influences the decision more than a technical list of vulnerabilities.

Also check the contract with the system's vendor. Some products offer paid extended support or fixes limited to critical flaws: ask in writing what is included and until when. If the vendor no longer exists, the plan must provide an alternative that does not depend on it, such as changing product or replacing the process. The plan should be reviewed when the vendor changes or when the replacement product is no longer available.

7. How to get started

This week, search your inventory straight away for systems the vendor no longer updates, and make a list with owner and function. If there is no inventory, start from a list of servers, workstations and equipment connected to the network. You do not need a complex project to begin: you need one line per system. The minimum line contains name, owner, function and end-of-support date, and is enough to get started.

Then choose the most exposed system or the one holding the most data, and complete the sequence of steps for it. The result will be a documented exception with a review date. Once the first case is closed, the model is easily repeated for the others. The second case will show you which parts of the model are truly general and which are specific to one application.

Frequently Asked Questions

Can I keep using an unsupported system if it is isolated?

You can use it as a temporary measure, not as a stable solution. Isolation greatly reduces the risk but does not eliminate it: a forgotten cable, a USB stick or remote access can bring it back onto the network. Record the decision as an exception with an owner and a review date, and plan the exit. Do not treat it as a normal system.

Is an antivirus enough to protect it?

No. An antivirus can intercept some known threats, but it does not fix the flaw the vendor no longer closes. On a system that cannot be updated, protection depends on other levers: isolation, access control, reducing active services and monitoring. Protection software can be part of the plan, but it is not the plan.

What changes if the system processes personal data?

The level of attention changes. The GDPR requires security appropriate to the risk, so a system holding personal data that cannot be updated requires documented measures and a risk assessment. In case of a breach, notification to the supervisory authority must be made within 72 hours of becoming aware of it, where required. Speak to your data protection officer or a consultant.

How long can I keep a system in this condition?

There is no universal duration. It depends on the system's impact, the possibility of replacing it and the residual risk. The practical rule is that every exception has a review date and an exit date, and that both are either respected or renegotiated through a written decision. An exception without a deadline is an oversight, not management.

What should I ask the software vendor?

Ask for the official end-of-support date, any available extensions and on what terms, the migration options, and whether fixes are provided only for critical flaws. Also ask for documentation on how to isolate the system correctly. Written answers let you plan, instead of hoping the problem does not surface.

About this content

General informational guide. Regulations and technical requirements change over time: always check the official sources and assess your specific case with a qualified technical or legal advisor.

GET IN TOUCH

Talk to our team.

We'll show how Altovar integrates into your infrastructure. No commitment, no sales pitch. Just a technical conversation.

BOOK A DEMO →