Protecting Upstream Operations When the Attacker Already Has a Foothold

    A multinational upstream oil & gas operator with drilling and production sites across the Middle East and North Africa

    63 views

    Challenge

    A routine IT penetration test revealed lateral movement paths into the OT network — Modbus-speaking PLCs controlling wellhead pressure systems were reachable from the corporate LAN.

    Solution

    FlintX deployed at the IT/OT boundary and within the DCS network — real-time Modbus and OPC-UA protocol analysis, network segmentation validation, and SOC integration via syslog.

    Results

    12 critical segmentation violations identified and remediated in the first week, 3 unauthorized remote access sessions detected and terminated, zero production downtime.

    The assumption was simple, and it was wrong: the OT network was air-gapped.

    For years, that belief had been the foundation of OT security strategy at a multinational upstream oil and gas operator with drilling and production sites spread across the Middle East and North Africa. The corporate IT team had their perimeter. The OT engineers had their control systems. And between them, supposedly, was nothing but air.

    Then a routine IT penetration test proved otherwise.

    The Pen Test That Changed Everything

    The IT security team had contracted a third-party firm to conduct an annual penetration test of the corporate network. Standard scope — email systems, Active Directory, VPN concentrators, the usual targets. But the pen testers were thorough. They followed lateral movement paths wherever they led.

    And one path led somewhere it should not have.

    Starting from a compromised engineering laptop on the corporate VLAN, the testers were able to traverse through a series of misconfigured firewall rules and reach a network segment that carried Modbus/TCP traffic. They could see polling data from PLCs that controlled wellhead pressure monitoring systems. They did not attempt to send write commands — the scope of the engagement did not permit it — but the implication was clear.

    If a pen tester could reach those PLCs, so could an adversary. And unlike the pen tester, an adversary would not stop at read-only.

    The Gap Between Policy and Reality

    On paper, the operator had a segmentation architecture that followed the Purdue Model. Level 3.5 had a demilitarized zone with firewalls separating IT from OT. Vendor remote access went through a jump server with multi-factor authentication.

    In practice, the architecture had eroded over time. A historian server that needed data from both networks had been given dual-homed NICs. A vendor had installed a cellular modem for "temporary" remote access during a commissioning project two years earlier — it was still active. Firewall rules had accumulated over dozens of change requests, and nobody had done a comprehensive review in over a year.

    The result was a network that looked segmented on the architecture diagram but was not segmented in reality.

    Deploying at the Boundary — and Beyond

    FlintX was deployed in two phases. The first phase placed sensors at every IT/OT boundary point — the firewalls, the historian server, the jump server. The goal was to validate whether the segmentation architecture actually matched what was documented.

    It did not.

    Within the first week, FlintX identified 12 critical segmentation violations:

    • 4 firewall rules that permitted broad traffic from corporate VLANs into OT segments
    • 2 dual-homed engineering workstations bridging IT and OT networks
    • 3 active VPN tunnels from vendor laptops directly into the DCS network, bypassing the jump server entirely
    • 1 cellular modem with a public IP address connected to a switch in the production control room
    • 2 OPC-UA servers exposing process data to segments that included general-purpose IT systems

    Each violation was a potential path for an attacker to move from IT into OT — exactly the kind of path the pen testers had found.

    Phase Two: Inside the DCS

    The second phase deployed FlintX sensors within the DCS network itself — monitoring Modbus/TCP traffic between the master station and field PLCs, OPC-UA communications between the DCS and the historian, and serial-to-Ethernet converters that bridged legacy 4-20mA instruments to the IP network.

    This phase was about establishing what "normal" looked like inside the OT environment. What are the expected polling intervals? Which PLCs should be communicating with which master stations? What does a legitimate firmware update look like versus an unauthorized one?

    Three weeks into the deployment, FlintX detected three unauthorized remote access sessions. Someone was connecting to the DCS network through one of the vendor VPN tunnels — outside of any scheduled maintenance window, from an IP address that did not match the vendor's known infrastructure.

    The sessions were terminated within minutes. Investigation revealed that the VPN credentials had been shared among multiple contractor employees, and at least one set had been compromised. The contractor could not confirm when or how.

    Building a Segmentation That Actually Works

    The remediation was not glamorous. It was firewall rule reviews, access control list updates, and decommissioning of unauthorized remote access points. FlintX provided the visibility needed to prioritize these changes — showing exactly which violations carried the most risk based on what they connected to.

    The cellular modem was removed. The dual-homed workstations were replaced with properly architected data diodes for one-way data transfer. The vendor VPN was reconfigured to require per-session approval with time-limited credentials.

    More importantly, FlintX continued to monitor the boundaries after remediation. Segmentation is not a one-time project — it is a continuous discipline. New devices get added. Firewall rules get modified. Vendors request access. Without continuous validation, the architecture will erode again.

    The Outcome

    Six months after deployment, the operator had achieved something that years of policy documents and architecture diagrams had not delivered: actual, verified, continuously monitored segmentation between IT and OT networks.

    Zero production downtime occurred during the entire deployment and remediation process. The OT engineers did not need to modify their control systems or change their operational procedures. The PLCs kept polling. The wellheads kept producing.

    What changed was that the operator stopped assuming its OT network was protected and started verifying it — continuously, passively, and with protocol-level awareness that generic IT security tools cannot provide.

    "We had an architecture diagram that said we were segmented. FlintX showed us what was actually happening on the wire. Those are two very different things." — VP of Digital Security

    Share: