TTechclick ⚡ XP 0% All lessons
Microsoft · Defender for IoT · Device InventoryInteractive · L1 / L2 / L3

OT Asset Discovery & Device Inventory — Microsoft Defender for IoT

Microsoft Defender for IoT builds a complete OT device inventory automatically — no agents, no active scanning, zero impact on PLCs or RTUs. This lesson covers how passive DPI captures every device attribute (vendor, model, firmware, OS, IP/MAC, protocols, Purdue level), how IT vs OT vs IoT classification works, how the network device map is built, and how the sensor surfaces unmanaged and rogue devices the CMDB never knew about.

📅 2026-06-18 · ⏱ 16 min · 5 infographics · live block demo · 🏷 10-Q assessment + AI Tutor inline

⚡ Quick Answer

Learn how Microsoft Defender for IoT passively discovers every OT, IoT and IT device — capturing vendor, model, firmware, IP, MAC, protocols and Purdue level — plus the network map and rogue-device detection.

🎯 By the end you will be able to

Read as:

Pick where you want to start

1

Passive discovery

SPAN/TAP, DPI, and zero OT impact.

2

Device attributes

Vendor, firmware, protocols, Purdue level.

3

Rogue devices

New-asset flags, unmanaged assets.

4

Inventory at scale

Cloud vs local, sites & zones, Sentinel.

🧠 Warm-up — 3 questions, no score

Just notice which ones make you pause. We answer all three inside the lesson.

1. How does Defender for IoT discover OT devices without disrupting them?

Answered in Passive discovery.

2. Which device type is a PLC or RTU classified as?

Answered in Device attributes.

3. How does the sensor detect a rogue laptop on the OT network?

Answered in Rogue devices.

Most engineers think…

Many people assume OT asset discovery means running an IP scanner or deploying agents on industrial controllers — which is exactly what you cannot do on live PLCs and RTUs.

Microsoft Defender for IoT takes the opposite approach: the OT network sensor sits passively on a SPAN mirror port or TAP and uses deep packet inspection (DPI) to extract device attributes directly from protocol traffic. The result is a continuously updated inventory with vendor, model, firmware, OS, IP/MAC, protocols, and Purdue level — with zero active scanning and zero agents on the OT devices themselves. Every device that communicates is automatically discovered, including rogue and unmanaged devices that the CMDB never registered.

① Passive discovery — how the sensor builds the inventory

The OT network sensor connects to a SPAN/mirror port or TAP on the OT network switch and captures a copy of all traffic — without ever injecting packets. This is the foundation of Defender for IoT's agentless approach: DPI extracts device attributes from the natural protocol exchanges that PLCs, RTUs, HMIs, and other OT devices already make.

Every device that sends or receives a packet is added to the inventory automatically. The sensor watches for protocol banners, device-specific handshakes (e.g. Siemens S7 identity requests, BACnet device descriptions), and MAC vendor OUIs to build the richest possible device record. Because no traffic is injected, there is zero operational impact on fragile OT devices — a core requirement for industrial environments where even a brief network hiccup can halt a production line.

A device record is created on first observation and enriched continuously as more protocol exchanges are seen. The sensor runs detection locally at the edge, so inventory is available even in air-gapped / locally-managed deployments where no cloud connection exists.

LegendSensor & flow arrowsPipeline stage labelsDiagram headingStage detail textCanvas panel
Figure 1 — How passive discovery builds a device record
The sensor captures protocol traffic via SPAN/TAP, runs DPI, and continuously enriches each device record in the inventory.How passive discovery builds a device recordSPAN / TAPcopy traffic passivelyDPI decodeextract attributesDevice recordcreated on first seenContinuous enrichfirmware, protocolsInventorycloud or local
The sensor captures protocol traffic via SPAN/TAP, runs DPI, and continuously enriches each device record in the inventory.
Quick check · Q1 of 10 · Understand

Why does passive SPAN/TAP capture cause zero operational impact on OT devices?

Correct: c. SPAN/TAP creates a copy of existing traffic. The sensor reads that copy passively and injects nothing onto the OT network, so PLCs, RTUs, and HMIs see no additional load or unexpected packets.
👉 So far: Passive SPAN/TAP + DPI = complete device discovery with zero agents and zero OT impact. Every communicating device is automatically added to the inventory.

② Device attributes & classification — what is captured

For each discovered device, Defender for IoT captures a rich attribute set: vendor and model (from MAC OUI and protocol headers), firmware version and OS (from DPI of management protocol exchanges), IP address and MAC address, open protocols observed (Modbus, DNP3, Siemens S7, EtherNet/IP/CIP, BACnet, OPC-DA/UA, IEC 60870-5-104, IEC 61850, Profinet, and more), and the auto-assigned Purdue level (0–5).

IT vs OT vs IoT classification

The sensor automatically classifies every device into one of three types. OT devices — PLCs, RTUs, DCS, SCADA HMIs, engineering workstations — are identified by industrial protocol use and OT vendor MAC OUIs. IoT devices — IP cameras, HVAC controllers, badge readers, printers — have a lighter protocol stack. IT devices — Windows/Linux servers, workstations, switches, Active Directory — use standard enterprise protocols. This three-way classification drives risk scoring, Purdue level placement, and alert prioritisation. The network device map visualises all nodes and their communication links, filterable by device type, Purdue level, subnet, or protocol — giving OT teams an interactive topology view of the entire OT network.

Figure 2 — Device attribute layers captured by Defender for IoT
DPI builds a device record from identity fields up through communication context.Device attribute layers captured by Defender for IoTIdentityVendor, model, IP, MAC addressSoftwareFirmware version, OS, open portsProtocol & PurdueObserved protocols, Purdue level (0–5), device type
DPI builds a device record from identity fields up through communication context.
Figure 3 — OT, IoT & IT classification — one sensor, three types
The OT network sensor classifies every communicating device into one of three types based on protocol and MAC OUI signatures.OT, IoT & IT classification — one sensor, three typesOT SensorDPI classifierOT — PLC/RTUOT — SCADA/HMIIoT — CameraIoT — HVAC/BadgeIT — Server/PCIT — Switch/Router
The OT network sensor classifies every communicating device into one of three types based on protocol and MAC OUI signatures.
📡
SPAN / mirror port
tap to flip

A switch feature that copies all traffic on monitored ports to the sensor — the passive capture mechanism that gives Defender for IoT full network visibility with zero OT impact.

🔬
Deep packet inspection
tap to flip

Parsing packet payloads beyond IP/TCP headers to extract application-layer data — vendor, model, firmware, protocol commands — used to build each device record.

🗺️
Network device map
tap to flip

The interactive topology visualisation in the sensor or Defender portal showing every discovered device as a node, with communication links, filterable by Purdue level, protocol, or device type.

🚨
New-asset alert
tap to flip

Triggered when the sensor sees a device it has never observed before — the primary mechanism for surfacing rogue and unmanaged devices in real time.

Name the attribute set

In an interview, listing all captured attributes — vendor, model, firmware, OS, IP, MAC, protocols, Purdue level, device type — shows you understand what makes passive DPI powerful. An active scanner can find IPs; DPI builds a full engineering profile of every device without touching it.

Quick check · Q2 of 10 · Remember

Which of the following is captured as a device attribute by the OT network sensor?

Correct: b. Defender for IoT captures vendor, model, firmware, OS, IP/MAC, open protocols, Purdue level, and device type (OT/IoT/IT) — all derived from passive DPI of protocol traffic, not from host logs.
👉 So far: Each device record captures: vendor, model, firmware, OS, IP/MAC, observed protocols, Purdue level (0–5), and device type (OT / IoT / IT). The network device map shows all nodes and their communication links.

③ Rogue & unmanaged devices — what the CMDB missed

Because discovery is purely passive and covers every device that communicates on the monitored segment, Defender for IoT detects every asset — including those that were never registered in the organisation's CMDB or asset management system. Any device seen for the first time is flagged as a New device in the inventory and can trigger a new-asset alert for SOC review.

Common rogue device scenarios uncovered by the inventory: an unauthorised contractor's laptop plugged into the OT switch at Level 1, a personal Wi-Fi access point connected behind an HMI, an unsanctioned IP camera sharing the OT VLAN, or a duplicate IP from a stale device that was replaced but never decommissioned. The network device map makes these visually obvious — a new node with no authorised communication baseline stands out as an anomaly.

Cross-Purdue-level communication anomalies — for example a Level 1 PLC talking directly to a Level 4 enterprise server, bypassing the IT/OT DMZ at Level 3.5 — are also surfaced through the topology view and alert engine, making segmentation violations visible as soon as they happen.

Figure 4 — Known asset vs rogue device — how Defender for IoT sees them
Authorised assets have a known baseline; rogue devices appear as new, unclassified nodes in the inventory and on the network map.Known asset vs rogue device — how Defender for IoT sees themAuthorised assetRegistered in CMDBKnown vendor, model, firmwareExpected communication peersPurdue level assigned atRogue / unmanagedFlagged as 'New device'Unknown or unexpected vendorCommunicates with OT controllersTriggers new-asset alert for SOC
Authorised assets have a known baseline; rogue devices appear as new, unclassified nodes in the inventory and on the network map.

Priya Nair, OT security engineer at Vidyut Power Pvt. Ltd. in Chennai, faces this

During a weekly Defender for IoT review, Priya sees a Windows laptop flagged as 'New device' at Purdue Level 1, communicating with two RTUs over Modbus. The CMDB has no record of this device.

Likely cause

A maintenance contractor plugged their personal laptop directly into the Level 1 OT switch without authorisation. The laptop is running Modbus polling software from a previous job.

Diagnosis

In the Microsoft Defender portal, Device inventory, filter by 'New' status and Purdue Level 1. The laptop has no asset tag, no authorised owner, and its communication peers are production RTUs. The network device map shows a new node with direct links to Level 1 controllers.

Defender portal ▸ Device inventory ▸ Filter: New + Level 1 ▸ Network map
Fix

Isolate the laptop from the OT switch immediately. Investigate with the maintenance team — confirm the contractor and uninstall the Modbus software. Register all authorised maintenance devices in the CMDB before they connect, and set a policy alert for any new device appearing at Purdue Levels 0–2.

Verify

Re-check the network map: the laptop node shows no active communication links. The new-asset alert is closed after the investigation is documented and the CMDB is updated with the new control.

'Rogue devices are rare' under-sell

In practice, OT environments almost always have undocumented devices — contractor laptops, old PLCs never removed, shadow IoT. The inventory's new-asset flagging is one of the most operationally valuable features. Never dismiss it as a corner case: in OT, unmanaged devices are the norm, not the exception.

▶ Watch a rogue laptop get flagged in the OT inventory

How Defender for IoT passively detects and surfaces an unregistered device. Press Play for the healthy discovery path, then Break it to see the detection gap.

① Laptop connectsAn unregistered contractor laptop is plugged into the Level 1 OT switch and begins sending Modbus poll requests to a PLC.
② SPAN capturesThe OT network sensor's SPAN port receives a copy of all traffic — the laptop's packets are included automatically, no configuration change needed.
③ DPI identifiesDeep packet inspection extracts the laptop's IP/MAC, sees Modbus traffic, assigns Purdue Level 1, and classifies it as an IT device on an OT segment.
④ Inventory flags itThe device appears in the inventory as 'New device' and a new-asset alert is raised in the Defender portal for the SOC to review.
Press Play to step through the rogue-device discovery path. Then press Break it.
Quick check · Q3 of 10 · Apply

A maintenance engineer plugs an unregistered laptop into an OT switch at Level 1. How does Defender for IoT surface it?

Correct: c. Every device that communicates is discovered passively. An unregistered device is flagged 'New device' in the inventory and triggers a new-asset alert for SOC review — regardless of whether it is an IT, OT, or IoT type.
👉 So far: Every new device the sensor has not seen before is flagged as 'New device' and triggers a new-asset alert — the primary mechanism for surfacing rogue and unmanaged assets in real time.

④ Managing inventory at scale — cloud, sites & zones

In cloud-connected mode, device inventory from all sensors is synchronised to the Microsoft Defender portal, giving a single-pane view across every OT site in the organisation. Analysts can filter, search, and export the full inventory from one place, and Microsoft Sentinel receives device-context data through the native data connector — so alerts automatically include asset owner, device type, firmware, and Purdue level without manual lookup.

Defender for IoT organises sensors under a hierarchy of sites and zones. A site maps to a physical location (e.g. a substation, a factory floor, a hospital wing); zones subdivide a site by network segment or process area. The inventory view can be scoped by site/zone, enabling large organisations — a utility with many substations or a manufacturer with multiple plants — to delegate inventory management to local OT teams while giving the central SOC a unified view.

Air-gapped deployments

In locally-managed (air-gapped) mode the device inventory lives entirely on the sensor or the on-premises management console — no cloud sync. OT teams access it through the sensor's local web interface. This suits highly regulated environments where no outbound cloud traffic is permitted, though it means each site's inventory is managed separately.

Figure 5 — Inventory sync — sensor to Defender portal to Sentinel
Cloud-connected sensors push device inventory to the Defender portal; Sentinel ingests asset context for richer OT alerts.Inventory sync — sensor to Defender portal to SentinelOT Sensordiscovers devicesDefender portalunified inventorySites & zonesscope by locationSentinel connectorasset context inalerts
Cloud-connected sensors push device inventory to the Defender portal; Sentinel ingests asset context for richer OT alerts.
Check inventory before responding to alerts

When a Defender for IoT alert fires, pull up the device record first — firmware, Purdue level, communication peers, last-seen time. An alert on a known, fully-patched switch means something different from the same alert on an unmanaged rogue device at Level 1. The inventory is your context layer, not just an asset list.

Quick check · Q4 of 10 · Analyze

An organisation has 12 OT sites that cannot connect to the cloud. How is device inventory managed in this scenario?

Correct: b. In locally-managed (air-gapped) mode the device inventory is stored on the sensor or the on-premises management console and accessed through the sensor's local web interface — no cloud sync is required.
👉 So far: Cloud-connected sensors sync inventory to the Defender portal, scoped by sites and zones. Air-gapped sites keep inventory local. Sentinel ingests device context to enrich OT alerts.

🤖 Ask the AI Tutor

Tap any question — instant, scoped to this lesson. No login, no waiting.

Pre-curated from vendor docs + community Q&A, scoped to this lesson. For a live prod issue, paste your export into chat.techclick.in.

📝 Wrap-up assessment — six more

You've answered 4 inline. Six left. 70% (7 of 10) marks the lesson complete on your profile. Tap Submit all answers at the end.

Q5 · Remember

Which capture method does the OT network sensor use to achieve passive, agentless discovery?

Correct: c. Defender for IoT uses a SPAN/mirror port or TAP to receive a copy of all OT network traffic passively. No packets are injected, no agents are installed, and OT devices see no additional load.
Q6 · Understand

A Siemens S7-300 PLC is discovered by the sensor. Which device type classification will it receive?

Correct: c. OT devices are identified by industrial protocol use (e.g. Siemens S7) and OT vendor MAC OUIs. A Siemens S7-300 PLC will be classified OT and assigned to an appropriate Purdue level automatically.
Q7 · Apply

You want to see which Level 1 PLCs are communicating directly with Level 4 enterprise servers in your OT environment. Where do you look?

Correct: b. The network device map in the sensor or Defender portal shows all discovered nodes and their communication links. Filtering by Purdue level makes cross-level connections — Level 1 to Level 4 — visually obvious as topology anomalies.
Q8 · Understand

What happens in the Defender for IoT inventory when the sensor observes a device it has never seen before?

Correct: b. New devices are flagged immediately in the inventory and can trigger a new-asset alert. This is the primary mechanism for surfacing rogue and unmanaged devices — no manual CMDB check required.
Q9 · Analyze

A large utility has 20 OT substations, some with no internet access. How should they manage device inventory across all sites?

Correct: d. Defender for IoT supports both cloud-connected and locally-managed (air-gapped) deployment modes. Cloud-connected sensors sync to the Defender portal; air-gapped sensors maintain inventory locally. The sites and zones hierarchy organises the inventory across both modes.
Q10 · Evaluate

Which statement best explains why passive DPI-based discovery is preferred over active scanning in OT environments?

Correct: a. OT devices are engineered for reliability in industrial processes, not for responding to IT-style network scans. Active scanning can cause PLCs and RTUs to fault, restart, or drop control loops. Passive DPI never injects traffic, eliminating this risk entirely.
Lesson complete — saved to your profile.
Almost! You need 70% (7 of 10) — re-read the path that tripped you up and tap "Try again".

🧠 In your own words

Type one sentence: what is the single most important reason Microsoft Defender for IoT uses passive DPI instead of active scanning for OT device discovery? Then compare with the expert version.

Expert version: OT devices — PLCs, RTUs, DCS — are engineered for reliability in industrial control loops, not for tolerating IT-style network probes; an active scan can cause a controller to fault, restart, or drop a control output, disrupting a live process. Passive DPI from a SPAN port or TAP reads a copy of traffic that is already flowing, extracts vendor, model, firmware, protocols, and Purdue level from natural protocol exchanges, and builds a complete engineering-grade inventory without ever injecting a single packet onto the OT network.

🗣 Teach a friend

Best way to lock it in — explain it in one line to a teammate. Tap to generate a paste-ready summary.

📖 Glossary

SPAN / mirror port
A switch feature that copies all traffic on monitored ports to a dedicated monitoring port — how the OT sensor receives a passive copy of all OT network traffic.
Deep packet inspection (DPI)
Parsing packet payloads beyond IP/TCP headers to extract application-layer data such as device vendor, model, firmware, and industrial protocol commands.
OT device
Operational technology device — PLC, RTU, DCS, SCADA HMI, engineering workstation — that directly controls or monitors industrial processes.
IoT device
Internet of Things device — IP camera, HVAC controller, badge reader — with a lighter protocol stack than OT controllers, classified separately by Defender for IoT.
Purdue model
A reference architecture for OT/ICS networks with Levels 0–5 defining device roles from field sensors (L0) to enterprise IT (L4/5), plus a Level 3.5 IT/OT DMZ.
Network device map
The interactive topology visualisation in the sensor or Defender portal showing all discovered devices as nodes with their communication links, filterable by Purdue level and device type.
New-asset alert
An alert triggered when the sensor observes a device it has never seen before — the primary mechanism for surfacing rogue and unmanaged devices.
Sites and zones
Defender for IoT's organisational hierarchy for grouping sensors and device inventory by physical site (location) and zone (network segment or process area).

📚 Sources

  1. Microsoft Learn — Microsoft Defender for IoT device inventory. learn.microsoft.com/en-us/azure/defender-for-iot/organizations/device-inventory
  2. Microsoft Learn — OT network sensor overview and passive discovery architecture. learn.microsoft.com/en-us/azure/defender-for-iot/organizations/architecture
  3. Microsoft Learn — Manage your OT device inventory from the Azure portal. learn.microsoft.com/en-us/azure/defender-for-iot/organizations/how-to-manage-device-inventory-for-organizations
  4. Microsoft Learn — Purdue reference model and OT network segmentation in Defender for IoT. learn.microsoft.com/en-us/azure/defender-for-iot
  5. Microsoft Learn — OT threat monitoring in enterprise SOCs — alert types and device context. learn.microsoft.com/en-us/azure/defender-for-iot/organizations/concept-supported-protocols
  6. Microsoft — Microsoft Defender for IoT product page. microsoft.com/en-us/security/business/iot/microsoft-defender-for-iot

What's next?

Got the inventory covered? Next, explore how Defender for IoT maps every device to the Purdue reference model (Levels 0–5), auto-detects cross-level segmentation violations, and enforces OT network segmentation policy.