A network segmentation diagram shows how a system is divided into zones of different trust, which firewalls sit between them, and exactly which ports are allowed across each boundary. This sample shows a classic DMZ / app / data layout plus a zero-trust path for remote staff. It is a topology diagram, not UML: UML covers software structure and behaviour; topology diagrams are the standard notation used alongside it for networks.
1. The Diagram
Read it left to right:
Traffic from the Internet (red, untrusted) can only reach the edge firewall, which also runs the WAF and DDoS protection. Only port 443 is open.
In the DMZ the load balancer terminates TLS. Nothing in the DMZ holds data; if it is compromised the attacker still faces a second firewall.
The internal firewall enforces zone ACLs: the load balancer may reach the web tier on 8080 and nothing else.
In the app zone the web tier talks to app services over mutual TLS. App services are the only hosts allowed into the data zone, on 5432 (PostgreSQL) and 8200 (Vault).
Outbound traffic (dashed) leaves only through the egress proxy with an allow-list, which blocks data exfiltration and malware call-backs.
Admins reach servers only via the bastion, whose sessions are recorded. Every zone ships logs to the SIEM.
Remote employees never get a network-level VPN. The ZTNA proxy checks identity, MFA and device posture, then forwards to one application at a time (green).
2. How to Read It
Real topology diagrams often use Cisco or cloud-vendor icons; here every device is a labelled shape so the diagram stays readable in any tool. The labels that matter most are the port numbers on the arrows: together they are the firewall rule set, and anything not drawn should be denied.
3. Variant: Zero-Trust Logical View
Zero trust replaces "inside the network = trusted" with per-request decisions. This variant uses the NIST SP 800-207 roles: a policy enforcement point in front of each resource asks a policy decision point, which combines identity, device posture and risk signals. Zones from the main diagram still exist, but they become a second layer rather than the main control.
4. Pros, Cons, and When to Use
Pros
Makes blast radius visible: you can see what an attacker reaches from each zone.
Port labels double as a reviewable firewall rule set.
Auditors (PCI DSS scoping, SOC 2) ask for exactly this picture.
Shows egress and admin paths, which are often forgotten.
Easy to compare with the live cloud security groups.
Cons
Drifts from reality unless generated from or checked against live config.
Hides identity-based controls that do not map to a network location.
Large estates need several diagrams at different zoom levels.
Says nothing about application-level threats such as injection.
When to use
Designing or reviewing VPCs, data-centre zones and firewall rules.
Compliance scoping: proving card or health data is isolated.
Planning a migration from VPN to zero-trust access.
When not to use
Explaining a login or API protocol (use a sequence diagram).
Data-level threat analysis inside one service (use a STRIDE DFD).
5. Diagrams-as-Code
Graphviz (DOT) version of the main topology, using clusters for zones: