Techbanq

Defensive security · control systems

Securing equipment that was built to obey, not to ask

A PLC executes any correctly formed instruction it receives, an HMI trusts its operator screen, and an engineering workstation can rewrite the logic of both. None of these can be hardened the way a server is, so their protection is built in the layer around them — who may reach them, over which protocol, saying what.

Serving Saudi Arabia from our Riyadh office and our engineering hubs in Lahore and Karachi

Market
Saudi Arabia
Written for
Plant Manager, CISO
Decision
Protect the control-system layer itself — PLCs, HMIs, historians, engineering workstations
Languages
EN · AR

Cybersecurity credentials

  • National Cybersecurity Authority, Saudi Arabia

    NCA registration

    Registered cybersecurity service provider

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

The components

What each piece of the control system cannot do for itself

Each component has a defence it inherently lacks, and the engineering task is to supply exactly that from outside it.

PLCs and RTUs

Cannot verify who sent an instruction, and may hang or fault when probed by tools written for IT equipment.

What the layer around it supplies: A conduit that admits only named stations, restricted to the command classes each legitimately sends.

HMIs and SCADA servers

Often run operating systems years past support, pinned by the vendor qualification of the control software above them.

What the layer around it supplies: Application allow-listing, removable-media control and isolation sized to the trust the OS no longer deserves.

Historians

Sit with one foot in the control network and one in the business network — the classic crossing point between the two.

What the layer around it supplies: A one-way publishing pattern: the business side reads a replica and no path leads back down to control.

Engineering workstations

Hold the project files and the software that can rewrite controller logic — the single most consequential machine on the estate.

What the layer around it supplies: Dedicated hardened machines, no dual use, tightly logged access, and project files versioned and backed up off the estate.

The protocols

Living safely with protocols that authenticate nothing

Modbus, and much of the protocol family a plant runs on, predates the idea of a hostile network: any device that can reach a controller and form a valid frame can read its registers or write to them. There is no password to strengthen and no certificate to install, and the equipment speaking these protocols will not be replaced for a decade or more, because it is qualified, paid for, and running a process that works. Treating that as a finding to remediate misreads the situation; it is a property of the layer, and the defence has to be designed assuming it.

Designing around it means the network does the authenticating the protocol cannot: reachability itself becomes the credential. Only the named HMI and the named engineering workstation can open a session to a given controller; protocol-aware inspection at the conduit distinguishes a read of a register from a write to it, and from a firmware download, and polices each differently; and the monitoring treats a write command from an unexpected source as the highest-severity event on the estate. The controller still trusts every well-formed frame — the design makes sure the only frames that arrive are ones that deserve the trust.

The estate

The inventory an ICS defence plan has to account for

A plan that names only the controllers misses most of the attack surface. This is the estate as an attacker would list it.

Field and control

  • PLCs
  • RTUs
  • IEDs
  • VFDs
  • Safety controllers

Supervision

  • SCADA servers
  • HMIs
  • Alarm servers
  • OPC servers

Data and engineering

  • Historians
  • Engineering workstations
  • Project files
  • Licence servers

Boundaries and access

  • OT firewalls
  • Jump servers
  • Data diodes
  • Vendor gateways

Our role

What we engineer around the controllers and what we leave untouched

  1. 01

    The layer, not the logic

    We design and implement the conduits, the allow-lists, the one-way patterns and the monitoring around the control system; the control logic itself, and any change to it, remains entirely the province of your engineers and your vendors.

  2. 02

    Every control tested against the process

    A restriction is proposed from observed traffic, run in monitor mode until it demonstrably blocks nothing legitimate, and enforced in a window with your operators watching the process indicators.

  3. 03

    Recovery treated as a control

    Controller programmes, HMI images and historian data get tested restore paths, because for equipment this fragile the honest last line of defence is the speed at which a known-good state comes back.

Which command paths are legitimate, and when a restriction may be enforced, are judgements only your control engineers can make; we bring the security engineering and put those judgements into force without touching what runs the process.

Common questions

What buyers ask

Our controllers cannot authenticate anything. Is replacing them the only real fix?+

No — and waiting for replacement is the worst plan, because the replacement cycle for qualified control equipment is measured in decades. The workable defence makes reachability do the protocol's missing work: if the only devices that can form a session with a controller are the named stations that legitimately command it, the missing authentication matters far less. That is buildable now, on the current equipment, with conduits, protocol-aware filtering and monitoring. When controllers are eventually refreshed, versions with secure protocol variants are worth specifying — but the network layer remains necessary even then, because a mixed estate is the reality for years either way.

Why do you treat the engineering workstation as more critical than the servers?+

Because of what it can change. A compromised HMI misleads an operator about one system; a compromised engineering workstation can silently rewrite the logic inside every controller its project files describe — the same capability the most serious industrial attacks on record were built on. It concentrates the project files, the vendor programming software and, too often, standing network paths to every PLC on the estate. That is why it gets the strictest treatment: a dedicated machine that does nothing else, no email or browsing, access logged per session, and its project files versioned so an unauthorised logic change is detectable by comparison rather than by symptom.

The historian only stores process values. Does it genuinely need this attention?+

Yes, but not mainly for its data — for its position. The historian is usually the one system with legitimate connections into both the control network and the business network, because the business side needs the production numbers. That makes it the natural crossing point for an attacker who has taken the corporate network and wants the plant. The defensible pattern publishes one way: the business side reads from a replica in a middle zone, and no session initiated from the business side can reach the control side at all. Its data also matters more than it first appears — process history is what incident investigation and regulatory reporting are reconstructed from — but position is the reason it sits high on the list.

Next step

Review your control layer with us

A component-level review of one control system that returns, for each fragile element, the specific compensating controls the network must provide.