Skip to content

OT Cybersecurity

Secure Remote Access to ICS - Security Principles

Remote access to OT networks step by step: a separate zone, MFA, session approval and recording, time-limited access and a cut-off procedure based on IEC 62443, NIST and CISA.

Józef SulwińskiJózef Sulwiński

Published Updated 24 min read

remote accessICSIEC 62443NIST 800-82zero trustVPN
Secure Remote Access to ICS - Security Principles
Contents
  1. Why remote access to OT needs its own principles
  2. Incidents in which remote access failed
  3. Zones and conduits for remote access
  4. 17 principles of secure remote access to OT
  5. How to assess a remote access solution
  6. Checklist for remote access to OT
  7. A few numbers
  8. Where we help
  9. Technical basis
  10. Related articles
  11. Sources

A vendor’s service technician connects to a controller at night because the line has stopped. A connection set up for one breakdown can stay open for years, with the same password. Anyone who learns that password has the same path to the process as the technician.

CISA advisories, the CERT Polska report and NIST guidance show that:

  • attackers enter OT networks through vulnerable edge devices and operator panels exposed to the internet, and then use default passwords,
  • CISA recommends removing OT connections from the public internet and running remote access through a VPN with phishing-resistant MFA,
  • NIST recommends initiating the connection from the OT side and removing access once it is no longer needed.

Why remote access to OT needs its own principles

A solution carried over from the office network, a VPN and a remote desktop with one password, gives the technician more than they need. An attacker who takes over that password gets just as much. According to NIST, OT environments differ from IT in several features that change how they are secured.

Feature IT OT
Priority Confidentiality and integrity of data come first. Human safety comes first, followed by protection of the process.
Downtime Momentary downtime is not a major risk, and a reboot is often acceptable. Even momentary downtime may be unacceptable, and outages are planned days or weeks in advance.
Software changes Patches are installed quickly, often automatically. Changes must be tested thoroughly and rolled out gradually, often by the system vendor.
Lifetime Components usually last three to five years. Components usually last 10 to 15 years, and sometimes longer.
Support Many vendors provide support. A single vendor often provides support.
Authentication Systems support standard login mechanisms. Some components support only a single authentication factor or none.

A controller without authentication executes a command from anyone who reaches it. Security measures therefore belong on the path to the controller: in the access gateway, in the jump server and in the firewalls between zones. This is one of the layers of defense in depth.

Incidents in which remote access failed

Incident What happened Lesson for remote access
Colonial Pipeline, May 2021 The attackers entered through an unused VPN profile with single-factor authentication and encrypted the IT systems. The company shut down the pipeline as a precaution because it could not rule out a compromise of the OT system. Remove unused profiles and accounts and require MFA, because an attack on IT was enough to stop OT.
Oldsmar, February 2021 The sodium hydroxide dosing setting was changed remotely from 100 to 11,100 ppm, and the operator reversed the change. The cause remains unexplained. Protect critical setpoints with value limits and a second confirmation, because an error and an intrusion have the same effect.
Attack on the Polish energy sector, December 2025 According to CISA, the attackers entered through vulnerable edge devices and reached HMIs and RTUs with default passwords. On more than 30 farms they damaged the RTU firmware, which cut the farms off from the grid operator. Patch edge devices and change factory passwords, because these opened the way to the controllers.
Volt Typhoon, advisory of February 2024 According to CISA, the group entered through vulnerable firewalls and VPN gateways and prepared movement from IT networks to OT. In some networks it kept access for at least five years. Treat edge devices as part of the control system and collect their logs centrally.
Pro-Russia hacktivist groups, advisory of December 2025 According to CISA, the groups searched the internet for VNC services and broke weak or default passwords. Through VNC they changed setpoints on HMI panels, disabled alarms and stopped devices. Remove HMI panels from the internet and limit remote accounts to view-only where changes are not needed.

In February 2021, the sodium hydroxide dosing setting at the water treatment plant in Oldsmar, Florida, was changed remotely from 100 to 11,100 ppm. The operator reversed the change. The Pinellas County Sheriff presented it as an intrusion. In 2023 the FBI stated that it could not confirm a targeted intrusion. The former city manager pointed to employee error. The FBI and the employee have not confirmed this account, so the cause remains unexplained.

According to CISA, weak passwords and an outdated operating system were likely used to gain access. CISA did not confirm the use of remote desktop software. The lesson for OT is the same in both versions: a remote setpoint change without a value limit and without a second confirmation is dangerous. Security measures must also work when an authorised person makes a mistake.

Zones and conduits for remote access

Remote access that lands directly in the control network bypasses every zone the company has designed. According to IEC 62443-3-2, the organisation groups assets into zones based on risk assessment. The standard recommends placing devices that connect to the system through external networks in a separate zone with its own requirements. It reasons that remote access lies outside the physical boundary of the system.

The connection between the remote access zone and the OT network is a conduit, a controlled communication channel between zones. Protect it like any other conduit. The target security level of a zone (SL-T, the level the zone is meant to achieve) follows from risk assessment. The standard prescribes no method for setting it and does not tie it to the type of connection.

NIST describes the DMZ, an intermediate network between the corporate network and OT, as a boundary where the organisation enforces traffic rules. IEC 62443-3-3 names a DMZ as an example of an effective boundary architecture. The article on OT network segmentation describes how to design zones.

17 principles of secure remote access to OT

Group Principles What it sets
Architecture 1-4 The group sets the path of the connection and who initiates it.
Authentication and privileges 5-8 The group sets who may enter, whose approval they need and with what rights.
Monitoring and records 9-12 The group sets what is visible during a session and what remains after it.
What to make available remotely 13-15 The group sets which systems are never made available, how files pass and how long access lasts.
Hardening and cut-off 16-17 The group sets how to protect access devices and how quickly to close access in an incident.

Architecture

Principle 1. Build access from two hops, with a different protocol and different credentials on each.

The first hop leads from the internet to the remote access zone, for example over a VPN. The second leads from that zone to the OT network through a jump server or a remote access gateway. A jump server is a computer through which a user may go further, and it is the only one connected to the OT network.

Do not run a single protocol, for example RDP, from the internet through the DMZ all the way to a workstation in the OT network. Why: one vulnerability in that protocol or one stolen password then opens the whole path.

Principle 2. Initiate the connection from the OT side.

The firewall blocks all incoming connections to the OT network, and a component in the OT network sets up the connection to the remote access zone itself. NIST lists initiating the connection from the OT environment among the practices for temporary access. Why: the OT network then has no port that constantly waits for a connection from outside.

Configure firewalls at zone boundaries to deny traffic by default and allow only explicitly permitted traffic. Record each permitted rule with its source address, destination address, port and justification. Why: a rule without a recorded reason stays in the firewall after the work ends.

Principle 3. Place intermediary servers in a DMZ with firewalls on both sides.

The jump server, the access gateway and the file scanning station sit in the DMZ. Do not store production data on servers in the DMZ. Why: a compromised server in the DMZ gives the attacker only an intermediary server, while the data and controllers stay behind the second firewall.

Configure firewall rules to allow traffic only between adjacent network levels. NIST gives an example: a device in the corporate network does not connect directly to a controller or an operator panel.

Principle 4. Remove every OT device from the internet.

This covers ports forwarded to controllers, HMI panels with a VNC service, cellular modems and VPN gateways running directly on OT devices. CISA recommends removing OT connections from the public internet. Why: in its December 2025 advisory, CISA describes groups that searched the internet for HMI panels with VNC and logged in with default passwords.

Check exposure regularly from outside with an attack surface scanner, for example Tenable One. Any OT service visible from the internet is a breach of this principle.

Protocol Port in the IANA registry Where it is used
Modbus TCP 502 The protocol is used in controllers, RTUs and sensors.
ISO-TSAP (S7comm) 102 The protocol is used in Siemens S7 controllers.
OPC UA 4840 The protocol is used in process data servers and newer controllers.
EtherNet/IP 44818 The protocol is used in controllers and drives.
BACnet 47808 The protocol is used in building automation.
DNP3 20000 The protocol is used in power and water utilities.
VNC (RFB) 5900 The service is used to view and operate operator panels at a distance.

Authentication and privileges

Principle 5. Require MFA for every remote access, preferably in a phishing-resistant form.

MFA (multi-factor authentication) requires at least two kinds of proof from the user, for example a password and a hardware key. CISA recommends phishing-resistant MFA for remote access to OT. NIST considers MFA an accepted good practice for remote access to OT applications.

Enforce MFA in the access gateway and on the jump server. Why: some OT components do not support multi-factor authentication, so the controller cannot check the second factor itself.

IEC 62443-3-3 assigns MFA for access over untrusted networks to SL 3, and the target level follows from risk assessment. The standard allows an organisation’s policy to require MFA for all remote access regardless of the level.

Method Phishing resistance Notes for OT
FIDO2 hardware key (WebAuthn) The method is resistant because it binds the response to the name of the service. Check that the access gateway and the client support WebAuthn.
Smart card with a certificate, e.g. PIV The method is resistant because it binds authentication to the TLS channel. The workstation the user connects from needs a card reader.
Hardware token with a one-time code (TOTP) The method is not resistant because the user enters the code manually. The token works without a phone, so it suits plants that do not allow phones.
Phone app with a code The method is not resistant because the user enters the code manually. The method requires a phone on site.
SMS code The method is not resistant, and NIST SP 800-63B-4 restricts its use. Do not use it for access to OT, because the code can be taken over after a SIM swap or number porting.

Principle 6. Require a person on the OT side to approve every session.

The technician submits an access request, the maintenance engineer approves it, and the gateway grants access for a set time. NIST recommends approving each remote maintenance session and informing operations staff about planned work. Why: the operator knows who is working on the installation and can stop the work when the process requires it.

At unattended sites, for example remote wind farms, a dispatch centre can approve the session. The approval can also come from a schedule of maintenance windows that someone approved in advance.

Principle 7. Give every remote user a separate account.

Assign each vendor account to a specific person and a specific work order. Do not use shared accounts such as “service_vendor1”. Why: a shared account makes it impossible to establish who made a change in the controller.

Disable unused accounts after a period set in the policy. The CIS Controls give 45 days of inactivity, and CISA recommends disabling dormant accounts when configuring remote access. Keep OT access accounts outside the corporate domain, or record a risk assessment for shared accounts. Why: NIST recommends weighing the risk that an account compromised in IT opens the way to OT.

Enable account lockout after several failed logins in the access gateway. Do not use it on operator workstations. Why: IEC 62443-3-3 warns that automatic lockout must not hinder an operator’s response in an emergency.

Principle 8. Give a remote user only the privileges the work order requires.

Separate the roles: reading controller parameters, changing the controller program, and starting and stopping the process. A technician diagnosing a fault gets read access only. Write access is granted for the duration of an approved maintenance window.

Protect critical setpoints regardless of who logs in. The controller or the SCADA system rejects values outside a physically plausible range, and every setpoint change triggers an alarm. A change to a critical setpoint requires approval by a second person.

IEC 62443-3-3 gives a change to a setpoint of a critical process as an example of an action that needs dual approval. The standard notes that dual approval is not used when people must be protected immediately, for example during an emergency shutdown.

Monitoring and records

Principle 9. Record and monitor all remote access traffic.

Each log entry contains a timestamp, the source (device, process or account), and the category, type, identifier and result of the event. An alarm should be raised for a connection at an unusual hour, from a new address, for a write to a controller and for a program change.

A firewall with deep packet inspection (DPI) reads addresses, ports and the content of industrial protocol commands. In Modbus TCP every command has a function code that tells whether the command reads or changes data.

Modbus function code What it does What to do
01-04 The commands read coils, discrete inputs and registers. Allow them from SCADA and HMIs and record them in the log.
05, 06, 15, 16 The commands write one or more coils or registers. Allow them only from engineering workstations in an approved maintenance window.
21, 22, 23 The commands write a file record, write a register with a mask, and read and write multiple registers at once. Treat them as writes, because a rule that blocks only codes 05, 06, 15 and 16 lets a data change through.
08 The command is used for diagnostics, and the specification reserves it for serial lines only. Block it in traffic from untrusted networks.
43 with sub-code 14 The command reads the device identification: vendor, product code and revision. Block it in traffic from untrusted networks, because it reveals data for looking up vulnerabilities.

Principle 10. Terminate the encrypted tunnel in the remote access zone and set up a separate session to OT.

The VPN tunnel ends in the DMZ, and a new session from the jump server leads to the OT network. Why: NIST points out that an intrusion detection system cannot assess encrypted traffic, so a tunnel to the controller itself hides commands from monitoring.

Disable split tunnelling in the VPN client, which is a simultaneous connection to the company network and to the internet. Why: a technician’s computer with a split tunnel connects the OT network to the internet.

Principle 11. Record sessions to critical systems and store the recordings out of reach of the people recorded.

A recording of the session screen and typed commands shows after an incident exactly what the user did. Store recordings in a separate system and protect them cryptographically against change. Why: a user with access to their own recordings can delete them.

Agree the scope of recording with a lawyer. Why: a recording captures people’s work, including vendor staff, and NIST recommends session auditing in consultation with legal counsel and in line with applicable law.

Principle 12. Give the operator a view of active sessions and the ability to end them.

The operator sees who is connected and to what, and can end a session with one action. A session ends by itself after a period of inactivity set by the policy. IEC 62443-3-3 assigns the capability to terminate a remote session to SL 2 and allows it to be limited to monitoring and maintenance sessions.

Before uploading software to a controller, extend the session and record the justification. Why: a disconnection during the upload interrupts the operation halfway.

What to make available remotely

Principle 13. Do not make functional safety systems available remotely.

This covers SIS, that is safety instrumented systems that shut the process down automatically. It also covers emergency shutdown, fire and gas detection and any system whose failure endangers life or the environment. Why: an error in the logic of such a system disables the last layer of protection for people.

According to IEC 62443-3-2, safety-related assets must sit in zones that are logically or physically separated from other zones. Change the logic of these systems on site, under the management of change (MOC) procedure.

Principle 14. Scan every file sent to the OT network in the DMZ.

Files with software, configuration and patches first go to a file scanning station in the DMZ. The station has no permanent connection to the OT network, scans files with several antivirus engines and records who sent each file. Why: NIST recommends malicious code protection at system entry points.

The station accepts only the file types the work needs, for example an export of a controller project. Files pass to OT through a controlled channel, and each transfer has an approved work order.

Principle 15. Enable access only for the duration of the work.

The account is activated on an approved request, works in a set time window and disables itself at the end. This model is called just-in-time (JIT) access. NIST recommends automatic disabling of temporary accounts and timers that remove remote access.

Why: a service account that remains active months after the work is a path nobody watches. At Colonial Pipeline, that path was a VPN profile nobody used any more.

Hardening and cut-off

Principle 16. Harden every device on the remote access path.

The jump server, the VPN gateway and the edge devices have only the functions needed for the job. Disable all other services, ports and programs. Why: every unneeded service is an additional vulnerability on a device visible from outside.

  • The jump server has two network cards, one to the DMZ and one to OT, with no forwarding between them.
  • On the jump server, clipboard sharing and file transfer are disabled unless the work order requires them.
  • The firewall on the server accepts connections only from the VPN gateway and sends them only to recorded addresses in the OT network.
  • Edge devices get patches first, because in the attacks described they were the point of entry.
  • Edge devices store no credentials and no domain administrator accounts, because CISA recommends removing them after the Volt Typhoon attacks.
  • Devices for which the vendor no longer issues patches are replaced, or separated and monitored more closely.
  • Logs from VPN gateways and firewalls go to one central location.

Base the IPsec VPN configuration on NIST SP 800-77r1. Do not use 3DES, MD5, SHA-1 or Diffie-Hellman groups 2 and 5. Why: NIST does not approve groups 2 and 5 and recommends AES with SHA-2 and group 14 or elliptic curve group 19, 20 or 21 instead.

Principle 17. Prepare and practise a procedure for cutting off remote access immediately.

According to NIST, incident handling includes preparation, detection and analysis, containment, eradication and recovery. The cut-off procedure has a recorded step for each of these stages. Why: during an incident nobody has time to work out who may switch off the VPN.

  1. The firewall has ready rules that block all remote access traffic and switch on with a single command.
  2. Remote access devices are labelled so that staff can disconnect them physically and quickly.
  3. A contact list states who decides on the cut-off, who carries it out and whom to notify, including vendors with active sessions.
  4. The procedure describes how to stop ongoing maintenance work safely and record the state of the installation.
  5. The team practises the procedure at the interval set in the response plan and measures the time from decision to full cut-off.

Disconnection from external networks must not stop the functions essential to the process. CISA also recommends practising manual operation and planning for OT devices that have stopped working.

How to assess a remote access solution

Assess a remote access solution against the same principles, regardless of vendor and price.

Question What to check
Does the OT side initiate the connection? Check that the firewall to the OT network can block all incoming connections.
Does phishing-resistant MFA work at every login? Check support for WebAuthn or certificate cards in the gateway and on the jump server.
Does the operator approve every session? Check that the gateway grants access only after approval and only for a set time.
Are sessions recorded? Check where the recordings are stored and who can delete them.
Does the solution read industrial protocols? Check that the rules distinguish reads from writes, for example Modbus function codes.
Do files pass through scanning? Check that file transfer to OT passes through a scanning station and a log.
Does the vendor hold an IEC 62443 certificate? Check the certificate: IEC 62443-4-1 covers the product development process, and IEC 62443-4-2 the requirements for a component. Conformity with IEC 62443-3-3 is assessed for the whole system.

Checklist for remote access to OT

  • Remote access passes through a separate remote access zone and a DMZ with firewalls on both sides.
  • No OT device is reachable directly from the internet, and exposure is checked regularly from outside.
  • The connection to the DMZ and the connection from the DMZ to OT use different protocols and different credentials.
  • Connections to the OT network are initiated from the OT side, and firewalls deny traffic by default.
  • Every remote access requires MFA, preferably phishing-resistant.
  • Every user has a separate account, and unused accounts are disabled.
  • A person on the OT side approves every session.
  • Service accounts work only in an approved time window.
  • Privileges separate reading, program changes and process control.
  • Critical setpoints have value limits, an alarm on change and a second approval.
  • All sessions are logged, and sessions to critical systems are recorded.
  • Alarms react to unusual hours, new addresses and writes to controllers.
  • The VPN tunnel ends in the DMZ, and traffic to OT is visible to monitoring.
  • Functional safety systems are not available remotely.
  • Files sent to OT pass through scanning in the DMZ.
  • Sessions end after inactivity, and the operator can end them manually.
  • Access devices are hardened and up to date, and devices without support are replaced.
  • The procedure for cutting off remote access is documented and practised.
  • Accounts, privileges and firewall rules are reviewed periodically.

A few numbers

Number What it means Source
55% This share of organisations has four or more remote access tools in their OT environment, according to a study of more than 50,000 devices. Claroty Team82, 2024
22% This share of organisations in the SANS survey had an ICS/OT incident, and half of these incidents began with external connectivity or remote access. SANS, 2025
10-15 years OT components usually last this long, and sometimes longer. NIST SP 800-82r3

Where we help

During an IEC 62443 audit, we check how remote access passes through zones and conduits and which of the principles described work in a specific installation. The result is a list of gaps with remediation priorities. An OT security assessment also covers a test of vendor access paths. A security architect designs the remote access architecture, and you are welcome to reach us through the contact form.

Technical basis

Principle / topic Where it is
IT and OT differences, component lifetime NIST SP 800-82r3, section 2.4 and Table 1
Zone for devices connecting through external networks IEC 62443-3-2:2020, ZCR 3.6 (recommendation)
Grouping into zones by risk, SL-T without a prescribed method IEC 62443-3-2:2020, ZCR 3.1 and ZCR 5.6
DMZ, rules between adjacent levels NIST SP 800-82r3, 5.2.3.1; IEC 62443-3-3, SR 5.2 (rationale)
1. Two hops NIST SP 800-53, SC-7
2. Initiation from the OT side, deny by default NIST SP 800-82r3, 6.2.10; IEC 62443-3-3, SR 5.2 RE 1 (SL 2+); NIST SP 800-53, SC-7(5)
3. DMZ NIST SP 800-53, SC-7 b; NIST SP 800-82r3, 5.2.3.1
4. Removing OT from the internet, ports CISA, Primary Mitigations (05.2025); CISA AA25-343A; IANA port registry
5. MFA NIST SP 800-53, IA-2(1), IA-2(2); NIST SP 800-82r3, 6.2.1.4.4; IEC 62443-3-3, SR 1.1 RE 2 (SL 3+) and SR 1.13; NIST SP 800-63B-4, 3.1.3.3 and 3.2.5
6. Session approval IEC 62443-3-3, SR 1.13 RE 1 (SL 2+); NIST SP 800-53, MA-4(5), AC-17(1); NIST SP 800-82r3, 6.2.10
7. Separate accounts, unused accounts, lockout IEC 62443-3-3, SR 1.1 RE 1 (SL 2+) and SR 1.11; NIST SP 800-53, IA-2, IA-8, AC-2(3); CIS Controls, Safeguard 5.3
8. Least privilege, dual approval IEC 62443-3-3, SR 2.1 RE 2 (SL 2+) and RE 4 (SL 4); NIST SP 800-53, AC-6
9. Logging and monitoring, Modbus codes IEC 62443-3-3, SR 2.8 and SR 6.2; NIST SP 800-53, SI-4; MODBUS Application Protocol V1.1b3
10. Tunnel ends in the DMZ, split tunnelling NIST SP 800-82r3, 5.2.3.3; NIST SP 800-53, SI-4(10), SC-7(7)
11. Session recording NIST SP 800-53, AU-14, AU-9(2), AU-9(3)
12. Session termination IEC 62443-3-3, SR 2.6 (SL 2+); NIST SP 800-53, SC-10, AC-12
13. Functional safety systems IEC 62443-3-2:2020, ZCR 3.3 (requirement); NIST SP 800-82r3, 4.2.2 and 5.3.1
14. File scanning NIST SP 800-53, SI-3, MA-4; NIST SP 800-82r3, 6.2.7 and 6.3.2.3
15. Access only for the duration of work NIST SP 800-53, AC-2(2); NIST SP 800-82r3, 6.2.10
16. Hardening, edge devices, IPsec NIST SP 800-53, CM-7, SI-2; CISA AA24-038A; NIST SP 800-77r1
17. Cut-off procedure NIST SP 800-53, IR-4, IR-8; IEC 62443-3-3, SR 5.2 RE 2 (SL 3+); NIST SP 800-82r3, 6.2.10 and 6.4; CISA, alert of 10.02.2026
Product certificates IEC 62443-4-1 (development process), IEC 62443-4-2 (components), IEC 62443-3-3 (system)

Sources

  1. CISA, FBI, EPA, DOE - Primary Mitigations to Reduce Cyber Threats to Operational Technology (6 May 2025) - CISA (accessed: )
  2. CISA AA25-343A - Pro-Russia Hacktivists Conduct Opportunistic Attacks Against US and Global Critical Infrastructure - CISA (accessed: )
  3. CERT Polska - Energy Sector Incident Report (29 December 2025), published 30 January 2026 - CERT Polska (accessed: )
  4. CISA - Poland Energy Sector Cyber Incident Highlights OT and ICS Security Gaps (10 February 2026) - CISA (accessed: )
  5. US Senate - hearing of 8 June 2021, testimony of J. Blount (Colonial Pipeline) - U.S. Government Publishing Office (accessed: )
  6. CISA AA24-038A - PRC State-Sponsored Actors Compromise and Maintain Persistent Access to U.S. Critical Infrastructure (Volt Typhoon) - CISA (accessed: )
  7. CISA AA21-042A - Compromise of U.S. Water Treatment Facility - CISA (accessed: )
  8. The Register - Florida water treatment plant hacked (9 February 2021) - The Register (accessed: )
  9. CyberScoop - Oldsmar water incident (10 April 2023) - CyberScoop (accessed: )
  10. Water Online - Former Official Claims Oldsmar Drinking Water Hack Was Really Operator Error (29 March 2023) - Water Online (accessed: )
  11. NIST SP 800-82 Rev. 3: Guide to Operational Technology (OT) Security (September 2023) - NIST (accessed: )
  12. NIST SP 800-53 Rev. 5 - OSCAL catalog, release 5.2.0 - NIST (accessed: )
  13. NIST SP 800-63B-4: Authentication and Authenticator Management - NIST (accessed: )
  14. NIST SP 800-77 Rev. 1: Guide to IPsec VPNs - NIST (accessed: )
  15. MODBUS Application Protocol Specification V1.1b3 - Modbus Organization (accessed: )
  16. IANA - Service Name and Transport Protocol Port Number Registry - IANA (accessed: )
  17. CIS Controls Assessment Specification - Control 5: Account Management - Center for Internet Security (accessed: )
  18. Claroty Team82 - 55% of OT Environments Contain Four or More Remote Access Tools (10 September 2024) - Claroty (accessed: )
  19. SANS - Why VPN and MFA Are Not Enough for OT: Evidence from the SANS State of ICS/OT Security Report (5 January 2026) - SANS Institute (accessed: )
  20. Tenable One - Attack Surface Management - Tenable (accessed: )

Need help in this area?

Our experts will help you assess the risk and plan next steps.

Talk to an expert

We'll discuss scope, methodology, and timeline.

Book a call