How Does Ransomware Jump to OT? Spread Paths and Cut-Off Points

Through which paths can a ransomware incident that starts in IT reach production? Dependency analysis for shared AD, the historian and remote access, and the countermeasures that can be taken for each path.
Technical summary
Technology
Active Directory, historian, MES, IDMZ, RDP, SMB
Topic
Ransomware spread paths from IT to OT and dependency analysis
Level
Intermediate–advanced
Prerequisites
  • Access to the plant’s IT/OT network diagram
  • Basic knowledge of the Active Directory structure
  • An operations representative who knows the IT dependencies of production processes
In this article8 sections

In most ransomware incidents that stop production, the attacker never touches the PLC. The attack starts on the corporate network; it reaches HMIs and SCADA servers through shared identity infrastructure, data transfer servers or remote access paths, or it encrypts only IT systems and stops the line indirectly. Below we go through the typical spread paths from IT to OT one by one, and for each give the cut-off point and a dependency analysis answering the question “if IT goes down, how long does production continue?”

Production stops even without a direct attack

In the 2021 Colonial Pipeline incident, ransomware affected corporate IT systems, and the company shut down pipeline operations as a precaution to contain the threat. This public example shows that production can stop without OT being encrypted. There are three typical reasons for stopping:

  • Dependency: If work orders, recipes, labels, quality approval or shipping information come from IT systems, then when these systems go down, the line can’t produce output even if it’s physically running.
  • Uncertainty: If it isn’t known whether OT has been affected, management chooses to stop production because of safety and quality risk.
  • Actual spread: Ransomware reaches Windows servers and clients in OT; HMI and SCADA screens are lost and the operator can’t see the process.

That’s why, from an industrial cybersecurity perspective, protection should answer two questions: Through which paths does ransomware reach OT? And even if it doesn’t, how dependent is production on IT?

Spread path: shared identity infrastructure

The most dangerous path is OT servers and clients being members of the corporate Active Directory. In human-operated ransomware operations, the attacker usually first gains domain administrator-level privileges, then distributes the payload across the whole domain through group policy (GPO), scheduled tasks or remote management tools. If OT computers are in the same domain, no distinction is made. They all go at once.

The same effect comes through other doors too: corporate administrator accounts being used in OT, the same local administrator password repeated across IT and OT, a single EDR or software distribution console managing all endpoints. If the central management console is compromised, the agents in OT can become the attacker’s tool.

To see the current state, list which domain OT computers are in and the contents of the corporate privileged groups:

# Computers in the OT OU (change the OU path to match your structure)
Get-ADComputer -SearchBase "OU=Production,DC=corp,DC=local" -Filter * `
  -Properties OperatingSystem, LastLogonDate |
  Select-Object Name, OperatingSystem, LastLogonDate

# Members of the corporate privileged groups
foreach ($g in "Domain Admins","Enterprise Admins","Administrators") {
  Get-ADGroupMember -Identity $g -Recursive |
    Select-Object @{n='Group';e={$g}}, Name, objectClass
}

Compare this output with the contents of the local Administrators group on OT computers. The goal is the complete list of corporate privileged accounts that can sign in to OT servers. This list often turns up the account of a contractor who left years ago or a help desk group added “temporarily.” Every account on the list is a vehicle for spreading to OT if compromised.

Cut-off point: A separate AD forest or domainless (workgroup) management for OT, not using corporate administrator accounts in OT, and separate management tools for OT. Tiering administrator accounts on the corporate side also makes spread harder; this approach is described in the Active Directory tier model article.

Spread path: the historian, MES and data transfer

Production data flowing to IT is a business necessity. The problem is how the flow is set up. Common risky designs:

  • A reporting server in IT making a direct SQL connection to the historian database in OT
  • An MES server sitting on the IT network and opening bidirectional connections to HMIs and PLCs in OT
  • Recipe, NC program or label files being pulled by OT clients from an SMB share in IT
  • Servers connected to both networks (dual-homed) acting as a bridge

SMB shares are especially dangerous: ransomware scans and encrypts the network shares it can reach, and all clients connected to that share are affected at the same time. One share, the whole line.

Cut-off point: All data exchange between IT and OT terminating in the IDMZ. One-way replication from OT to a copy in the IDMZ for the historian, a controlled transfer server in the IDMZ for file transfer, and no communication passing directly from IT to OT is the basic rule. The network design of this separation is covered in the IT/OT segmentation guide.

Spread path: the jump server and remote access

A jump server set up for access to OT is a ready-made gateway for an attacker if designed wrong. A jump server that is a member of the corporate domain, accessed via RDP with corporate accounts and with broad access to the OT network becomes OT’s front door the moment the domain falls. Remote desktop tools installed on HMIs and vendors’ permanent VPN tunnels also open paths that bypass IT entirely.

Cut-off point: The jump server being in a separate identity infrastructure, MFA on access, access to OT only with defined targets and protocols, and disabling RDP features such as clipboard copying and drive redirection. To move vendor connections to this model, see the machine builder remote access article.

Spread path: shared infrastructure services

Less visible dependencies are just as dangerous: the corporate backup server that backs up OT servers, the corporate WSUS that distributes updates to OT, a virtualization cluster shared by IT and OT, a single hypervisor management console. In attacks that encrypt the virtualization layer, OT virtual machines running on the same cluster are affected too. If the hypervisor falls, everything on it falls.

Cut-off point: The OT virtual infrastructure being in a separate cluster and on a separate management network; OT backups being kept in an immutable store that can’t be reached with corporate identities. For separation on the backup side, the identity isolation in the backup infrastructure article gives practical steps.

IT dependency analysis: how long can production run without IT?

Cutting the spread paths doesn’t mean production won’t be affected by an IT outage. That’s why a dependency analysis is done for each line or plant. Run the analysis in a workshop with operations, maintenance, quality and logistics representatives; for each IT service, ask “what happens if this service isn’t working tomorrow morning?” The table below is an example working format; fill in the durations for your plant.

IT service Function in production Manual alternative How long it can hold out
ERP / work orders What to produce A printed production plan The period the plan covers
MES Recipes, traceability A local recipe copy, paper forms Determined per plant
Label / barcode server Product and pallet labels A stock of pre-printed labels As long as the stock lasts
License server SCADA/HMI software license A grace period, depending on the vendor Verify from the vendor documentation
AD, DNS, NTP Sign-in, name resolution, time Local services in OT Uninterrupted if there’s a local service

The license server row is the one most often skipped. Some SCADA and HMI software validates licenses over the network and may stop working after a certain time when the connection is lost. Get written confirmation of this behavior from the vendor for each product. The output of the analysis becomes an input to the OT section of the business continuity plan; from a compliance standpoint, the plan’s structure is described in the IT/OT business continuity in production article.

Cut-off points and prioritization table

The table below puts each spread path side by side with its cut-off point and implementation difficulty. The prioritization principle is simple: start with the control that closes the widest blast radius and touches production the least.

Spread path Cut-off point Difficulty Priority
Shared AD Separate OT identity infrastructure High 1
Direct connection from IT to OT Flows terminating in the IDMZ Intermediate 1
Remote access / jump server Separate identity, MFA, narrow rules Intermediate 1
SMB shares IDMZ transfer server Low to medium 2
Shared virtualization / backup Separate cluster, immutable backup High 2
IT service dependency Local services, manual procedure Varies 2

Although AD separation is difficult, it comes first, because it prevents a single identity breach from taking down all of OT. If the migration will take a long time, an interim step can be applied: with a group policy linked to the OT computers’ OU, add the corporate privileged groups to the “Deny log on locally”, “Deny log on through Remote Desktop Services” and “Deny access to this computer from the network” rights; define separate administrator accounts used only in OT. Try the restriction on a single test station first. If service accounts and engineering software aren’t affected, roll it out; if they are, resolve the dependency first.

Recovery is outside the scope of this guide; for preparing backups and the recovery process, see the recovery readiness against ransomware guide, and for protecting OT backups, the backup security and ransomware readiness service. Verifying the spread paths on site and turning the cut-off points into a roadmap are among the first steps of OT/ICS and industrial cybersecurity work. For general measures, CISA’s #StopRansomware resources are also a reference point.

Frequently asked questions

If OT systems don’t use Windows, is there no ransomware risk?

There is. PLCs and field devices are generally not ransomware’s direct target; but most of the HMIs, SCADA servers, historians and engineering workstations that manage them run on Windows. Production can also stop without OT being touched at all, because it depends on IT systems such as labeling, work orders and shipping.

Is it mandatory to remove OT servers from the corporate domain?

It isn’t a mandatory rule, but it’s one of the most effective cut-off points. OT servers joined to the corporate domain can be affected in a single move when a domain administrator identity is compromised. Alternatively, a separate forest for OT, with no trust or a limited one-way trust, and separate administrator accounts can be used.

Is it right to cut the IT/OT connection when ransomware is detected?

It’s usually the right first step, but you need to know in advance which services production depends on IT for. Define the cut-off decision and who makes it in advance in the incident response plan; verify with an exercise which manual methods production will continue with when the connection is cut.

How is historian data transferred safely to IT?

Set up one-way replication from the OT historian to a copy in the IDMZ and limit IT users to accessing only this copy. That way no direct connection from IT to the historian in OT is needed. Where high assurance is required, hardware-based one-way transfer such as a data diode can also be considered.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Cybersecurity

Firewall, VPN, endpoint security and threat prevention. We reduce the attack surface and measure and report the risks.

Knowledge Center

Other articles