Securing a Connected City's Infrastructure Before the Threat Becomes Physical

    A European metropolitan authority managing smart building automation, tunnel ventilation, and intelligent traffic control across a city of 1.8 million residents

    48 views

    Challenge

    Rapid IoT/OT expansion created an attack surface no one fully understood — BACnet-enabled HVAC controllers, tunnel SCADA systems, and traffic signal controllers shared network segments with minimal segmentation.

    Solution

    FlintX provided unified visibility across heterogeneous OT protocols (BACnet, Modbus, SNMP), asset classification by criticality, and continuous monitoring tuned to operational context.

    Results

    1,200+ connected assets mapped across 47 sites, 8 misconfigured controllers with public-facing interfaces identified and isolated, incident response time reduced by 70%.

    Smart cities are built on a promise: connect everything, and the city gets more efficient, more responsive, more livable. But there is a gap between the promise and the reality, and it is a gap that keeps growing. The more you connect, the more you expose.

    A European metropolitan authority learned this the hard way. Not through a breach — they were lucky enough to avoid that — but through a growing realization that they had built a connected city without fully understanding what they had connected.

    The Expanding Attack Surface Nobody Mapped

    The city had been modernizing its infrastructure for nearly a decade. Building automation systems managed HVAC and lighting across 23 municipal buildings. Tunnel ventilation and fire suppression systems were controlled by SCADA platforms in 6 road tunnels. Intelligent traffic control coordinated 180+ signalized intersections. Water treatment and distribution had its own set of PLCs and RTUs.

    Each system had been deployed by a different vendor, at a different time, with a different integration partner. The building automation used BACnet. The tunnel SCADA ran on Modbus/TCP. The traffic controllers communicated over NTCIP and SNMP. The water systems used a mix of DNP3 and proprietary protocols.

    What they had in common was that they were all, to varying degrees, connected to the city's municipal network. And nobody had a complete picture of what that network actually looked like.

    The city's IT department managed the corporate network — email, file servers, employee workstations. The OT systems were managed by the respective engineering departments, each with their own contractors and maintenance schedules. There was no single team, no single tool, and no single view of the entire connected infrastructure.

    The Incident That Was Not an Incident — But Could Have Been

    The catalyst for action came not from a cyberattack, but from a maintenance error. A contractor updating firmware on a BACnet controller in a municipal office building accidentally pushed a configuration change that propagated to controllers in two other buildings on the same network segment. HVAC systems in those buildings switched to full heating mode during a summer heatwave.

    It took 14 hours to diagnose the root cause. Not because the problem was technically complex, but because nobody could see the relationship between the devices. The contractor did not know that the buildings shared a network segment. The building management team did not know that a firmware update on one controller could affect another. The IT department did not monitor BACnet traffic at all.

    Nobody was hurt. But city leadership recognized that if a maintenance error could cascade across buildings undetected for 14 hours, a deliberate attack could do far worse — and might not be caught at all.

    Building Visibility Across Heterogeneous Systems

    FlintX was deployed across the city's infrastructure in phases, starting with the highest-risk systems: tunnel ventilation and traffic control.

    The deployment model was the same at each site — passive sensors connected to SPAN ports on network switches, mirroring traffic without introducing any inline components. No changes to the SCADA systems. No agents on PLCs. No modifications to vendor configurations.

    The challenge was protocol diversity. Unlike a factory or a power plant where you might deal with one or two industrial protocols, the city's infrastructure spoke half a dozen. FlintX's protocol-aware engine parsed BACnet, Modbus/TCP, SNMP, DNP3, and several vendor-specific protocols simultaneously, building a unified asset inventory and communication map across all of them.

    Within the first two weeks of deployment across 47 sites, FlintX had identified 1,247 connected assets. The city's existing asset records accounted for approximately 800.

    What the Gaps Revealed

    The 400+ previously unknown assets were not all surprises. Some were network switches and media converters that the IT team knew existed but had never catalogued. Some were test equipment left connected after commissioning. But eight of them were genuinely alarming:

    • 3 BACnet controllers in municipal buildings with interfaces accessible from the public internet via misconfigured NAT rules
    • 2 tunnel ventilation PLCs with default vendor credentials still active
    • 1 traffic signal controller with an open Telnet port on the municipal network
    • 2 SNMP-managed switches in traffic cabinets configured with the default "public" community string, allowing anyone on the network to read — and in one case, write — device configurations

    None of these had been exploited. But each one represented an open door that the city did not know existed.

    Tuning Detection to Operational Context

    One of the challenges unique to smart city deployments is the diversity of operational contexts. A BACnet write command to an HVAC controller at 3 AM might be perfectly normal during a scheduled maintenance window and deeply suspicious on any other night. A traffic signal controller rebooting during rush hour is an operational emergency; the same reboot at 2 AM during a planned update is routine.

    FlintX's alerting was tuned to these contexts. Maintenance schedules were integrated so that expected vendor activity would not generate false positives. Criticality classifications ensured that anomalies on tunnel safety systems were escalated immediately, while a misbehaving thermostat in a storage room was logged but not treated as urgent.

    This context-aware approach reduced alert noise by over 80% compared to the city's initial pilot with a generic network monitoring tool — which had generated so many alerts that the operations team stopped checking them within two weeks.

    Reducing Response Time From Hours to Minutes

    Before FlintX, the city's incident response for OT issues was essentially manual. Someone would notice something wrong — a tunnel fan not responding, a building too hot, a traffic signal stuck — and call the responsible engineering team. That team would dispatch someone to the site. The technician would connect a laptop to the local network and start diagnosing.

    Average time from detection to diagnosis: 6-8 hours. And that assumed someone noticed the problem in the first place.

    With FlintX providing continuous monitoring and automated alerting, the city established a centralized OT security operations workflow. Anomalies are detected in real time, contextualized with asset and protocol information, and routed to the appropriate response team with enough detail to begin diagnosis remotely.

    Incident response time dropped by 70%. Not because the city hired more people, but because the people they had could finally see what was happening.

    "We built a smart city. FlintX helped us understand what we had actually built — and showed us the gaps between our ambition and our security reality." — Chief Digital Officer

    Share: