GUIDE · ZERO TRUST NETWORK

Network segmentation: reducing the attack surface with VLANs and rules

This guide explains how to divide the corporate network into zones using VLANs and rules, reducing the damage a single compromised device can cause. By the end you will know which prerequisites to check, how to design the zones, the order in which to apply changes and how to confirm that the separation really works.

Updated october 2026
8 min read
Altovar Technical Editorial Team

1. Why segmentation reduces the damage

A flat network is one in which every device can talk to almost every other. That works as long as everything is running normally, but if one computer is compromised, an attacker can move on to other systems without meeting any obstacles. Segmentation creates predictable obstacles: it divides the network into zones and defines which flows are permitted. The damage from an incident therefore stays confined to the zone where it occurred, and becomes easier to detect.

One point that often causes confusion needs clarifying. A VLAN, or virtual local area network, separates traffic at the local network level, but on its own it is not a security control. Traffic between one VLAN and another passes through the router or firewall, and that is where the rules must be written. A VLAN without rules is just a label on the switch ports: it tidies the network, but does not protect it.

2. Prerequisites: inventory and flows

Before creating a single VLAN you need to know what is on the network. An inventory of systems, devices and shared services is the starting point. Without it you risk isolating a print server or an authentication service that users depend on, and work grinds to a halt. Also check dependencies: many applications need directory, name resolution or update services that must remain reachable from the zones that use them.

The second prerequisite is a flow map: who must talk to whom, on which ports and in which direction. A perfectly detailed level is not needed to begin. A readable list that application owners can check is enough, showing the purpose and reason for each flow. What is not documented tends to stay open forever, because nobody knows whether it can be closed.

  • Up-to-date inventory of systems, devices and printers.
  • List of shared services: directory, files, email, backup, updates.
  • An owner for each application that crosses several zones.
  • Administrative access to the routers, switches and firewall.
  • A change window agreed with the people who will use the systems.

3. Designing the zones: grouping systems and people

Good segmentation starts from a few clear groups, not from dozens of micro-zones that are hard to manage. A common model distinguishes the user network, the server network, the network for devices and connected equipment, the guest network and a zone dedicated to managing the network itself. Each zone has a different level of trust, and the rules between zones reflect that difference: traffic from users to servers flows only on the necessary ports, while traffic from guests to the internal network should not pass at all.

The table summarises the purpose of each zone. Adapt it to your organisation: some firms have no physical servers, others have many devices. What matters is that every system ends up in a specific zone and that every zone has an owner. If a system does not clearly fit any zone, it usually means it is not yet clear what it is for, and that is the point to resolve before moving any cable.

ZoneWhat it containsUnderlying rule
UsersStaff workstations and laptopsReaches servers only on the ports of the services used
ServersArchives, management software, internal email, directoryResponds only to authorised requests
DevicesPrinters, sensors, connected equipmentNo direct access from the internet, minimal traffic to the reference servers
GuestsVisitor Wi-Fi and personal devicesOutbound internet only, no access to the internal network
ManagementRouters, switches, firewall and administration panelsReachable only by dedicated administrative accounts

4. Operational steps: the implementation sequence

The work must be done in order, because each step makes the next one safe. The greatest risk is changing the network in a single intervention without knowing which dependencies will break. Proceed with one group of systems at a time, and allow time to notice any blockages before moving on. Record each blockage with its root cause, because the same cause tends to recur in other zones.

During the first implementation it is advisable to run an observation phase: the necessary flows are allowed, the rest is blocked but denials are logged. Only after a period agreed with the owners do you move to permanent blocking. In this way, excessive caution is corrected before it stops work. During this period, application owners must be able to report a missing flow immediately.

  1. List the systems and group them by function and risk level.
  2. Define the zones and assign an owner to each one.
  3. For each zone, list the flows required towards other zones and towards the internet.
  4. Create one VLAN per zone and assign the switch ports and Wi-Fi profiles.
  5. Write the firewall rules: first the permitted flows, defined specifically, then block everything else.
  6. Move systems one group at a time and verify that the applications work.
  7. Enable logging of denials and review the rules at the end of the observation period.

5. Common mistakes

The most frequent mistake is the generic rule: allowing everything between two zones because one application is not working, and then forgetting about it. A broad rule cancels the benefit of segmentation and is rarely narrowed, because nobody wants to risk going back to the problem. When a flow does not work, find out which port or address is missing and open only that, with a description and an owner.

Other mistakes are just as concrete: leaving device administration interfaces in the user zone, placing printers and sensors in the same zone as servers, forgetting that the backup server must be reachable only from authorised zones. Another classic is isolating a system without checking that name and directory services remain available: the symptom looks like an application fault, but the cause is in the network.

6. How to check that segmentation works

Segmentation must be tested, not assumed. From a workstation in each zone, try to reach the systems it should not see: if the connection succeeds, the rule is too broad. Also test the flows that must work, because an unintended block is as serious as excessive openness for those who work with those systems every day. Keep the results of these tests in a dated document.

Then review the denial logs over a period of time. Blocked traffic that recurs regularly almost always indicates an application that has not been mapped, whereas an isolated denial may be an attempt that deserves attention. Ask your supplier to show you where to find these logs and who receives alerts. A check performed only once is not enough: it must be repeated with every network change.

7. How to start

The first useful step is small. Choose a zone that does not touch critical systems, for example the guest network or the device network if they are already physically separated, and define minimum rules for it. Document the result, then move on to the next zone. Every completed zone becomes a template for the rest, because the same questions about flows almost always recur.

This week, ask your supplier to produce the flow map for the most important system, with the list of ports and owners. With that document in hand you can decide which zone to isolate first, without improvising during an outage. The document remains useful even when the supplier or staff change, because it describes the network without depending on anyone's memory.

Frequently Asked Questions

Do I need new switches to create VLANs?

Not always. Many networks already support VLANs, but this must be verified on each device. If a switch does not support them, the issue becomes replacing it rather than designing the segmentation. Ask your supplier for an inventory of devices with their supported features, because this affects costs and timelines.

Does segmentation alone stop an attack?

No. It reduces the chance that an incident spreads, but it does not prevent the initial attack. It must be combined with strong authentication, system updates, verified backups and log monitoring. Its value lies in limiting the damage and making unusual movement between different zones easier to spot.

Can I segment without stopping work?

Yes, if you proceed group by group and include an observation phase. The risk of blockage almost always comes from unmapped flows: a name service, a printer, an application calling an update server. With a flow map, changes can be made at low-impact times.

How much does segmentation cost?

It depends on the number of systems, the devices already in place, any component replacements and the design time. There are no reliable general figures that apply to every organisation. Ask your supplier for a proposal that separates the inventory and design work from any hardware supply.

Who should manage the rules over time?

One person or group with clear responsibility. Rules change along with applications and people, and without an owner exceptions accumulate. The supplier can handle the technical side, but the decision on a new flow must be approved by whoever is accountable for the system involved.

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 →