Abstracts

Modern software rarely fails in isolation, rather at the boundaries: sensor in- and outputs, network transitions, at geographic borders, and in the gaps between what a system does and what a user perceives.

This talk presents a collection of real-world testing scenarios drawn from complex customer projects executed on real hardware in a controlled cloud device lab, exploring error patterns such as overlapping time windows, split system state, cached assumptions, boundary instability that often survive traditional automations.

The talk advocates for a deliberate balance of automation and manual testing, supported by real devices and controlled environments, to expose the failure modes that matter most in production.