Enforced Once, Drifted Forever: Why Segmentation Policies Stop Matching the Network
A segmentation policy is accurate on the day it is applied and immediately starts going stale, because devices get patched, decommissioned, reassigned, and moved, and every one of those changes can put a policy out of sync with what the network actually looks like. New devices show up without anyone updating the rules that were written for the old ones, firmware updates change how a device behaves on the wire, and temporary exceptions granted for a single project quietly become permanent. Most programs have no reliable way to know when that drift has happened, so audits become manual exercises, often driven by spreadsheets and tribal knowledge, and confidence in existing policy erodes over time until nobody is quite sure whether the segmentation map matches reality. This session examines drift as the quiet failure mode of segmentation: not a dramatic breach, but a slow accumulation of small mismatches that eventually leaves critical assets less isolated than anyone realizes. We'll also look at the operational cost of policy sprawl on switches and firewalls that were never built to hold thousands of granular rules, and what happens to performance and manageability as that sprawl grows. From there, we'll walk through what continuous validation looks like in practice across IoT, OT, IoMT, and IT environments, including how teams can detect drift as it happens rather than discovering it during an incident or an audit. Attendees will leave with a practical way to check whether their own segmentation is still doing what they think it is, along with questions they can bring back to their own teams to start that conversation.