When factories, buildings and networks are connected, network location is no longer a reason to trust a request. What zero trust asks of operational environments.
QDT Editorial··2 min read
Traditional security drew a perimeter and trusted what was inside it. Connected operations break that assumption. A plant has contractors on site, gateways reaching the cloud, vendors with remote access and devices that cannot be patched on schedule. Zero trust is the response: stop treating network location as evidence of trustworthiness.
What zero trust says
NIST Special Publication 800-207 describes zero trust as an approach in which no request is trusted by default because of where it comes from. Access to each resource is decided per request, using identity, device state and context, with the least privilege necessary. It is a set of principles, not a product.
Identify
User, device or service
↓→
Assess
Health, context, risk
↓→
Authorise
Least privilege for this resource
↓→
Monitor
Log, re-check, revoke
FIG.Every request is evaluated; network location alone grants nothing.
Why operational environments are different
Information technology can usually install an agent, enforce multi-factor authentication and patch monthly. Operational technology often cannot: controllers run for years on fixed firmware, availability can matter more than confidentiality, and a security control that interrupts a process is itself a risk. Zero trust in these settings is applied around devices as much as on them.
Practical building blocks
01
Enterprise and cloud
Identity provider, ERP, analytics
02
Demilitarised zone
Brokered, inspected data exchange
03
Site operations
Servers, historians, gateways
04
Devices and controllers
Authenticated, patched where possible
FIG.Segmenting a connected operation. Compromise of one zone should not reach the next.
Know the estate. You cannot protect devices you have not inventoried. Passive discovery helps where active scanning is unsafe.
Segment. Group devices into zones and control the conduits between them, in line with the zones-and-conduits model in IEC 62443. A compromise in one zone should not reach the next.
Give everything an identity. Users, services and devices should authenticate individually, with credentials that can be rotated and revoked.
Broker remote access. Vendor and engineer access should be time-limited, logged and routed through a controlled path, not a standing VPN.
Monitor for the unusual. Baselines of normal device behaviour make deviations visible.
Plan recovery. Backups of configurations and tested restoration matter as much as prevention.
Start where the risk is
A full programme is a multi-year effort. A realistic start is to map the highest-consequence assets, close the most exposed remote paths, and introduce segmentation where a failure would hurt most. Each step should be reversible and tested in a window that operations accepts.
How QDT approaches it
QDT's security products and managed services cover cyber-security, managed services and telecom security, and our IoT implementations treat device identity and update paths as part of the architecture. Security posture depends on the environment and the scope of work; we describe controls and practices rather than guarantees.
Instrumenting a building is easy. Closing the loop between what sensors see and what operators do is where energy, maintenance and comfort actually change.
Connecting machines is the visible part of industrial IoT. Understanding how the plant actually runs is the part that determines whether the data is useful.
15 January 2026 · 3 min read
Start a conversation
Discuss the question behind the article.
Every article starts from a practical operating question. Tell us yours.