Learn
A2.2.4 | CONCEPTS AND APPLICATIONS OF NETWORK SEGMENTATION
01 | ONE NETWORK DOES NOT NEED ONE SHARED SPACE
Imagine a school where visitor laptops, student computers, staff devices and networked cameras all share one flat network. A device searching for a local address can generate traffic that reaches many unrelated devices. A compromised visitor device may also have an unnecessarily broad set of reachable targets.
Network segmentation divides a network into smaller sections, physically or logically. The sections can organise traffic and resources, while controlled connections between them can enforce access policy. The aim is to give devices the communication they need without making every device equally reachable.
For example, students may need the learning service and printers, while guests need only external internet access. Staff records and camera management should have more restricted paths. Segmentation provides boundaries where those differences can be implemented.
This lesson explains segmenting, subnetting and VLANs. These ideas work together, but they are not three interchangeable names for the same thing.
02 | THREE IDEAS WITH DIFFERENT JOBS
| Idea | What it describes | School example |
|---|---|---|
| Segmentation | The broader design of separate network sections and their boundaries | Separate guest, student, staff and service areas |
| Subnetting | Dividing an IP address block into smaller prefix-defined networks | Allocate distinct IPv4 subnets to student and staff groups |
| VLANs | Logical Layer 2 LAN separation on switching infrastructure | Place selected ports in VLAN 10 or VLAN 20 |
A common design assigns one IP subnet to each VLAN and routes between them where required. The VLAN supplies local link separation; the subnet supplies IP addressing structure; routing and policy determine permitted communication between sections.
Physical separation can use separate equipment and links. Logical separation can share hardware while keeping traffic in distinct logical groups. Sharing a switch does not necessarily mean sharing one broadcast domain.
03 | UNDERSTAND THE TRAFFIC BEFORE DIVIDING IT
An Ethernet frame carries data across a local link. A unicast frame targets one link-layer destination. A broadcast frame targets all participants in its Layer 2 broadcast domain, subject to the network’s forwarding rules.
A broadcast domain is the area over which those Layer 2 broadcasts are propagated. An ordinary switch floods a broadcast to other eligible ports in that VLAN, not back out its incoming port. The switch does not normally forward that same Layer 2 broadcast into a different VLAN.
For example, IPv4 ARP uses a broadcast to ask which local device has a particular IP address. Devices that do not own that address still receive the request at the link layer and decide whether it is relevant. A larger shared domain can therefore create more unwanted traffic and processing.
Do not assume a switch sends every unicast frame to everyone. It normally forwards a known destination selectively; unknown unicast and other traffic have their own handling. The activity later isolates broadcast behaviour so the boundary is easy to see.
04 | VLAN: SEPARATE LOGICAL LANS ON SHARED SWITCHES
A virtual local area network (VLAN) defines a logical Layer 2 LAN, normally a separate broadcast domain, on switching infrastructure. A managed switch can place one set of ports in a student VLAN and another set in a staff VLAN.
A student frame is not simply bridged into staff ports because both cables use the same physical switch. The VLAN membership controls which logical LAN the frame belongs to. This permits different groups to share infrastructure while keeping their local traffic separate.
A VLAN can also extend across multiple switches using suitable links and configuration. Students on different floors can belong to the same VLAN, while staff and students sitting beside one another can belong to different VLANs. A VLAN is not defined by a classroom wall or geographical distance.
VLAN IDs label logical membership. The number 20 does not automatically create subnet 192.168.20.0/24; administrators choose and configure the addressing separately. A tidy numbering convention is helpful documentation, not a protocol requirement.
05 | ACCESS PORTS, TRUNKS AND SHARED INFRASTRUCTURE
A typical access port attaches an endpoint to one configured VLAN. The endpoint’s ordinary data frames are commonly untagged on that access link; the switch associates them with the port’s VLAN. There are specialised exceptions, such as voice arrangements, but this is the core learning model.
A trunk link can carry traffic for multiple VLANs between suitable devices. IEEE 802.1Q tagging identifies VLAN membership for tagged Ethernet frames. Each carried VLAN remains logically distinct; sharing the cable does not merge their broadcast domains.
A trunk needs correct configuration at both ends, including the VLANs permitted to cross it. Some configurations use an untagged native VLAN. You do not need vendor commands here, but you should understand why simply connecting two switches does not guarantee the intended VLAN separation.
Only required VLANs should be carried to the places that need them. Configuration errors can create unintended exposure or broken connectivity, and a shared uplink or switch can remain a common performance or failure dependency.
ONE PHYSICAL MANAGED SWITCH
VLAN 10 | STUDENTS
Student A and Student B
VLAN 20 | STAFF
Staff A and Staff B
VLAN 30 | GUESTS
Guest A and Guest B
Local Layer 2 forwarding stays within the relevant VLAN. Inter-VLAN communication requires routing.
06 | SUBNETTING: ORGANISE THE IP ADDRESS SPACE
An IP address alone does not tell a host which addresses belong to its local subnet. A prefix length, or an equivalent IPv4 subnet mask, identifies the network portion. The remaining bits identify positions within that subnet.
For a gentle IPv4 example, 192.168.10.0/24 uses the first 24 bits as the prefix. Its mask is 255.255.255.0. Addresses in that block range from 192.168.10.0 to 192.168.10.255. In an ordinary /24 LAN, .0 is the network address and .255 is the directed broadcast address, leaving .1 to .254 for interface addresses.
A host uses its address and prefix to decide whether a destination is on-link. For an ordinary remote-subnet destination, it sends toward an appropriate router, often its default gateway. The host does not send the remote server’s Ethernet frame directly across every network.
These are private teaching addresses, not instructions to change a real school network. The next IP addressing lesson develops address types further; here the purpose is to understand how subnets create manageable IP sections.
07 | DIVIDE ONE /24 INTO TWO /25 SUBNETS
Suppose 192.168.10.0/24 must be divided into two equal address blocks. Extending the prefix by one bit gives two /25 subnets. The new mask is 255.255.255.128.
The first block spans .0 to .127 and the second spans .128 to .255. Each contains 128 total addresses. In these ordinary IPv4 LAN examples, removing the network and broadcast addresses leaves 126 possible interface addresses per subnet; gateways also consume interface addresses.
This is a division of address space, not a bandwidth upgrade. It does not by itself configure switch VLANs or enforce a security policy. Those corresponding network and routing settings must also be implemented.
| Subnet | Ordinary interface range | Directed broadcast | Possible gateway, if configured |
|---|---|---|---|
| 192.168.10.0/25 | 192.168.10.1–192.168.10.126 | 192.168.10.127 | 192.168.10.1 |
| 192.168.10.128/25 | 192.168.10.129–192.168.10.254 | 192.168.10.255 | 192.168.10.129 |
Notice the boundary: .126 and .129 are in different /25 subnets even though their first three decimal numbers match. Compare the prefix, not a rule such as “same first three numbers always means same subnet”.
08 | VLAN AND SUBNET: CONNECT THE TWO LAYERS
A common plan pairs VLAN 10 with one IP subnet and VLAN 20 with another. Devices within the same VLAN and IP subnet can communicate locally, subject to host controls. Devices in different VLANs need a routing function for ordinary IP communication between them.
The VLAN is the Layer 2 grouping; the subnet is the Layer 3 addressing structure. These are coordinated choices rather than an identity. Putting different IP subnets on one flat VLAN does not provide the same Layer 2 separation as configuring separate VLANs.
Similarly, assigning addresses from the same subnet to devices in isolated VLANs does not magically connect them. Ordinary local address resolution will not cross the VLAN boundary. Such an inconsistent design can break communication.
For the rest of this lesson, assume one subnet per VLAN, correct host configuration and no special bridging or overlay exceptions. This keeps the separation and routing behaviour clear.
09 | ROUTING ENABLES A PATH; POLICY CONTROLS ITS USE
A router, a Layer 3 switch or a suitable firewall can route between subnets and VLANs. A default gateway is the local router address a host uses when no more specific route applies. Enabling routing creates a possible path; it does not necessarily restrict which applications or destinations may use it.
An access control list (ACL) or firewall policy can permit required traffic and deny unnecessary traffic at the boundary. For example, students may access the learning server over HTTPS while guests are denied access to internal services.
The difference matters: VLAN separation alone is not a complete security policy. If unrestricted routing connects all VLANs, many cross-group targets can still be reachable. Conversely, a boundary without any required exceptions may prevent legitimate printing, name lookup or learning-service access.
Apply least privilege and test both directions and relevant responses. Stateful firewalls can track permitted connections, while ordinary ACL behaviour depends on its rules and placement. Network permissions also do not replace the application’s own authentication and authorisation.
10 | SEE WHERE A BROADCAST GOES
Send one student device’s Layer 2 broadcast in the example below. In the flat arrangement, all other listed endpoints share its broadcast domain. With VLAN separation, only the other listed student endpoint receives it.
The count covers the six displayed endpoints only. It omits infrastructure interfaces, frame copies on trunks and responses. It is a scope comparison, not a measured throughput claim. The router does not forward this original Layer 2 broadcast into the other VLANs.
11 | PERFORMANCE: LESS UNNECESSARY TRAFFIC, NOT MAGIC SPEED
Containing broadcasts can reduce irrelevant traffic delivered to unrelated groups. This can save link capacity and endpoint processing, especially where large or noisy domains generate substantial local traffic. Smaller sections can also make a fault or traffic storm easier to locate and contain, within the limits of shared infrastructure.
Segmentation can support efficient resource management: administrators can monitor demand by group, place services near their users, allocate address space and apply suitable controls. Traffic prioritisation or rate limits can be configured where supported, but VLAN membership alone does not reserve bandwidth.
Important bottlenecks remain. Student and staff VLANs may share a physical uplink, a firewall or one internet connection. A saturated shared connection can slow both groups even though their broadcasts are separate. Routing or inspection devices also need sufficient capacity.
Full-duplex switched Ethernet avoids normal Ethernet collisions already. Creating VLANs principally separates broadcast domains; do not claim it makes those links collision-free for the first time. Congestion, collisions and broadcast scope describe different issues.
12 | SECURITY: LIMIT THE REACH OF A PROBLEM
Segmentation can reduce the set of targets a compromised device can reach. A guest laptop should not need direct network access to staff records. An IoT camera may need to contact a recording service but not student laptops. Appropriate boundary policy can permit the required service while blocking unrelated paths.
This can limit lateral movement: an attacker’s attempts to move from one compromised part of a system to other targets. It does not guarantee that compromise cannot spread. Allowed services may have vulnerabilities, devices within one segment may still communicate, and stolen administrator access can undermine the design.
Use layered protection. Maintain endpoint security, application access checks, patched infrastructure, protected administration and monitoring. Restrict and validate trunk configurations rather than assuming a VLAN label is encryption or proof of identity.
A VLAN does not encrypt traffic. Devices sharing a VLAN are not automatically trustworthy, and traffic inside it may need additional controls. In particular, a guest VLAN may still need client isolation to prevent guest-to-guest access; separation from staff is a different requirement.
13 | APPLY A SIMPLE INTER-SEGMENT POLICY
Our school model has student, staff and guest VLANs plus a service VLAN. The configured routing policy permits students and staff to reach the learning service over HTTPS; it denies guests that internal access. It denies cross-user-group traffic. Within a VLAN, ordinary local switching is allowed in this simplified model.
Choose a source and destination. The model distinguishes a local path from permitted or denied routed traffic. It assumes healthy links, correct addressing and a working service. An allowed network path does not prove the user may view every page: the application must still check permissions.
Only this service and the displayed paths are modelled. Internet access, DNS exceptions, printer policy and stateful return traffic would need their own defined rules in a real design.
14 | PRACTICAL APPLICATIONS AND SHARED SERVICES
| Setting | Possible segmentation | Why and what still needs planning |
|---|---|---|
| School | Student, staff, guest and service VLANs/subnets | Limit unrelated traffic and internal access; permit required learning, printing and DNS paths |
| Hospital | Clinical, administrative, guest and medical-device sections | Reduce unnecessary reachability; define safe access to essential services and test availability |
| Business | Finance, general users, servers and visitors | Apply role-specific paths and monitor demand; shared uplinks still need capacity |
| Industrial site | Operational devices separated from office endpoints | Restrict paths to control systems; carefully plan updates, monitoring and recovery |
A shared DHCP service is often outside a client’s VLAN. Since the initial broadcast does not simply cross a router, a DHCP relay can forward the relevant exchange to the service according to configuration. Each subnet needs an appropriate allocation scope or pool.
Shared DNS, printing and application services also need deliberate routes and permissions. Segmentation succeeds when required work continues while unnecessary communication is reduced, not when every section is disconnected from everything else.
15 | PLAN, IMPLEMENT AND TEST THE BOUNDARIES
Begin with a communication map: which users and devices need which services? Group devices by role and risk, allocate suitable address blocks, and decide where VLAN membership, routing and policy will be enforced. Document gateways, required services and shared dependencies.
Then test a small controlled implementation. Check that permitted traffic succeeds, forbidden paths fail, address allocation works and monitoring records useful evidence. Test from actual endpoint locations; a diagram alone cannot verify a switch port or firewall rule.
Plan for growth and failure. Leave appropriate address capacity, size routed links and inspection devices, protect administration and test recovery. Avoid so many tiny sections that policy becomes unmanageable, or such large groups that the intended boundary disappears.
Segmentation adds configuration and support work. A wrong port assignment can place a device in an unintended group; a missing route or rule can break legitimate tasks. Review changes and maintain documentation rather than treating the initial setup as finished forever.
16 | CHECK THE CONCEPT IN CONTEXT
Choose the most directly relevant idea for each situation. Some questions focus on a boundary, others on addressing or enforcement. A complete design often needs all three, so read the exact function being asked about.
17 | COMMON CONFUSIONS TO RESOLVE
| Claim | More accurate explanation |
|---|---|
| A subnet is the same as a VLAN | A subnet is an IP addressing section; a VLAN is a logical Layer 2 LAN |
| Different VLANs can never communicate | Routing can connect them; policy should control the permitted paths |
| Changing IP prefixes alone gives strong isolation | Address planning alone does not enforce all required link or security boundaries |
| A VLAN guarantees faster internet | Shared bottlenecks and demand still determine performance |
| VLAN traffic is encrypted | VLAN separation and encryption are different mechanisms |
| Guest isolation is complete once staff are in another VLAN | Guest-to-guest paths and required services need separate consideration |
Use the layer distinction to keep the ideas straight: VLAN membership controls the logical LAN, subnet prefixes describe IP networks, routing enables movement between networks, and policy controls what movement is allowed.
18 | EXPLAIN A MECHANISM AND ITS CONSEQUENCE
For explain, go beyond naming features. State how a mechanism works and connect it to the outcome using because, therefore, so that or resulting in.
For performance: “Placing student and staff ports in separate VLANs creates separate broadcast domains. A student’s Layer 2 broadcast is therefore not flooded to staff access ports, reducing irrelevant traffic delivered to staff endpoints. The groups can still share an uplink, so segmentation does not remove every bottleneck.”
For security: “Separate VLANs and subnets provide boundaries between guest and staff devices. A firewall policy can deny guest-to-staff traffic while allowing required services, limiting unnecessary reachability. This does not replace host security or application access checks.”
For resource management: “Subnetting allocates distinct address blocks to groups. Combined with correctly configured VLANs and gateways, this makes allocation, monitoring and policy easier to organise. Required shared services must still be reachable through defined routes and permissions.”