6 Network Micro Segmentation Examples From Real Deployments
Real network micro segmentation examples include ADT isolating life-safety alarm systems, QBE Insurance protecting hybrid cloud applications, Ixom ring-fencing legacy manufacturing apps, St. Luke’s University Health Network and Main Line Health securing medical devices, and Andelyn Biosciences enforcing segmentation policies across a pharmaceutical plant. Each limits how far an attacker can move once inside.
Micro segmentation splits a network into small zones, often down to individual workloads or devices, and allows only the traffic each one needs. All six examples below come from case studies published by the segmentation vendors involved (Illumio and Elisity). Figures are vendor-reported, not independent audits, so treat them as indicative.
What These Network Micro Segmentation Examples Have in Common
Three patterns run through every deployment. Each started with visibility, mapping real traffic before writing rules. Each protected a specific high-value or fragile group of systems first rather than segmenting everything at once. And each used a monitor or test mode before enforcing policies, so legitimate traffic was not broken.
6 Real Network Micro Segmentation Examples
1. ADT: Protecting Life-Safety Alarm Systems Without Re-Architecting
ADT, the US home and business security provider, runs systems that handle alarm signals where delays can affect safety. According to Illumio’s case study, ADT used host-based micro segmentation to isolate these life-critical systems and reduce the risk of an attacker spreading into them, without redesigning the underlying network. The lesson: when downtime is unacceptable, software-based segmentation can be layered on existing infrastructure.
2. QBE Insurance: Hybrid Multi-Cloud Segmentation for Zero Trust
QBE, an Australian-headquartered global insurer, runs applications across data centers and multiple clouds. Illumio reports that QBE used micro segmentation as part of its zero trust program, applying consistent policies across those environments. The lesson: a single policy model across on-premises and cloud avoids gaps between separate tools.
3. Ixom: Ring-Fencing Legacy Applications in Manufacturing
Ixom, an industrial chemicals supplier in Australia and New Zealand, depended on legacy applications that could not easily be patched or replaced. Illumio’s case study says Ixom deployed its segmentation agents in two days and used them to ring-fence those older systems. The lesson: segmentation is often the most practical protection for systems that cannot be upgraded.
4. St. Luke’s University Health Network: 85,000 Devices in 46 Days
St. Luke’s, a US health network, needed to protect a large estate of clinical and connected devices. Elisity reports that it brought segmentation to about 85,000 devices across 15 hospitals in 46 days, using identity-based policies enforced through existing network switches. The lesson: hospitals with many devices that cannot run agents can segment through the network layer instead.
5. Main Line Health: Securing Biomedical Devices Across 150 Sites
Main Line Health, a US health system with about 150 facilities, holds more than 100,000 IoT, operational technology and medical devices, according to Elisity. The vendor reports deployment began within three days. The lesson: in healthcare, starting with device visibility and grouping devices by function makes policies manageable at scale.
6. Andelyn Biosciences: 2,700 Policies in 90 Days
Andelyn Biosciences, a US gene therapy manufacturer, had a converged IT and manufacturing network and an earlier segmentation project that had stalled with another product. Elisity reports that Andelyn put about 2,700 policies in place within 90 days. The lesson: a stalled project is not proof segmentation cannot work; the approach and tooling matter.
Comparison Table: Approach, Scale, Timeline and Outcome
| Organization | Sector | Approach | Scale or timeline (vendor-reported) | Main goal |
|---|---|---|---|---|
| ADT | Security services | Host-based | Not stated | Isolate life-safety systems |
| QBE Insurance | Insurance | Host-based, hybrid cloud | Not stated | Zero trust across clouds |
| Ixom | Manufacturing | Host-based | Agents deployed in 2 days | Protect legacy apps |
| St. Luke’s | Healthcare | Identity-based via network | 85,000 devices, 15 hospitals, 46 days | Secure clinical devices |
| Main Line Health | Healthcare | Identity-based via network | 150 sites, 100,000+ devices, 3 days to deploy | Secure medical and IoT devices |
| Andelyn Biosciences | Life sciences | Identity-based via network | 2,700 policies in 90 days | Contain lateral movement |
Lessons Learned From Real Deployments
The examples split into two technical approaches, and the choice drives speed and coverage.
Host-based (agent) segmentation, used by ADT, QBE and Ixom, installs software on servers and workloads. It suits data centers and clouds where you control the operating system, and it applies the same policy wherever the workload runs.
Network or identity-based segmentation, used by the three healthcare and life sciences organizations, enforces rules through switches and identity data without agents. It suits environments full of devices that cannot run software, such as infusion pumps, imaging systems and factory controllers.
Picture an illustrative regional manufacturer with a modern ERP system in the cloud and a factory floor of older controllers. Applying agents everywhere would fail on the factory side; network-only controls would miss cloud workloads. A split plan works better: agents for servers and cloud, network-based enforcement for operational technology. Use this decision rule: for each group of assets, ask whether you can install and maintain software on it. If yes, host-based segmentation gives finer control; if no, enforce at the network or identity layer.
The second lesson is sequencing. Every example started with one high-risk group, proved the policies in monitor mode and then expanded. Projects that try to segment everything at once tend to stall.
Common Mistakes These Deployments Avoided
Reading the case studies side by side also shows what not to do. None of the six began by writing rules from architecture diagrams, which are often out of date; they observed real traffic first. None tried to enforce on day one, because a single wrong rule can stop a clinical system or a production line. And none treated segmentation as a one-off project. Applications change, new devices arrive and old rules become too permissive, so each organization needed an owner and a regular review. Teams that skip these steps usually end up with either broken services or rules so broad they block nothing.
How to Apply These Examples to Your Network
- Pick your crown jewels: the systems whose compromise would hurt most.
- Map traffic for those systems over a few weeks.
- Choose host-based, network-based or both using the rule above.
- Write allow rules in monitor mode and review what would be blocked.
- Enforce gradually, starting with the most sensitive group.
- Expand and review policies as applications change.
In the US, HIPAA expects safeguards for health systems, and in the UK the NHS Data Security and Protection Toolkit asks suppliers and trusts to protect networks; segmentation supports both.
FAQs
St. Luke's University Health Network segmented about 85,000 devices across 15 hospitals, according to its vendor's case study.
No. They come from vendor-published case studies and should be treated as indicative.
Not always. Host-based tools use agents, while network and identity-based tools enforce through existing infrastructure.
Healthcare, financial services, manufacturing and critical services appear often because they protect sensitive or fragile systems.
It varies with scope. The examples above report first results in days to months, usually starting with one group of systems.
Yes. It limits access between systems, which supports the zero trust principle of granting only what each connection needs.
Planning Your Own Deployment
These network micro segmentation examples show that starting small, with clear visibility, beats trying to segment everything at once. If you want help choosing the right approach for your network, Cyber Security Solutions Ltd offers a free, no-obligation assessment through our contact page.
