OT Segmentation Under Operational Constraints

The talk argues that OT segmentation projects usually fail for operational reasons rather than missing firewall features. Narrow maintenance windows, fragile legacy systems, vendor-managed support contracts, and operators who put uptime and safety first all shape what can be changed. The speaker ranks priorities as safety, then operations, then cyber, then compliance, and warns that an IT-led firewall rollout that causes an outage can cost the project access for months or years.

The recommended approach starts with observation, not enforcement. Teams capture traffic over a long enough period to see cyclical flows, design zones and conduits, and begin in monitor-only mode at a low-risk canary site. Deep packet inspection can later restrict specific actions such as reads versus writes, but it is best added in stages. Rollback should be planned and pre-approved, with alarm-based triggers that operators can invoke themselves, and the plan should be validated against real process behavior.

The talk also covers drift. Temporary exceptions for vendor access tend to become permanent, and firewall rules gradually erode toward a flat network. Periodic configuration diffs, treating unapproved security-control changes as incidents, and assigning a durable owner help counter this. In one engagement, a six-week monitor-first period that gave operators better visibility turned them into internal champions for later sites.