#1Why convergence is now unavoidable
Predictive maintenance, real-time quality control, dynamic capacity planning — every modern manufacturing initiative depends on data leaving the shop floor and crossing into enterprise systems. The conversation has shifted from 'should we connect?' to 'how do we connect safely?'
#2The Purdue model — still useful, with updates
The Purdue enterprise reference architecture remains a useful mental model, but most real factories now collapse some of its zones and add new ones (edge compute, cloud telemetry). We modernise it rather than discard it: zones still exist, but each one publishes intentional data products to the next.
#3Security must lead, not follow
OT systems were designed for availability, not security. Bringing them into reachable networks means assuming compromise and limiting blast radius — segmentation, identity at the device, secure boot, signed firmware updates, and continuous monitoring tuned for OT signatures rather than IT ones.
#4A data platform built for the shop floor
Shop-floor data is dense, high-frequency, and operationally critical. We invest in a unified namespace at the edge, stream it to a time-series platform in the cloud, and expose curated marts back to operators on the floor — closing the loop in seconds instead of overnight.

