A quiet perimeter device proves nothing unless it can show it is healthy
In perimeter security, "no alarm" only means something if the device can prove it is healthy. Five health signals a dependable device should report.
Embedded Engineering Blog
Practical notes from the engineers at Gandeevan on firmware, Embedded Linux, IoT and perimeter security: the problems that show up in the field, and how we think about them.
An A/B update does not roll back a bad image that hangs
A/B update schemes roll back a bad image only if the device reboots. How U-Boot bootcount works, why a hang defeats it, and the watchdog settings that close the gap.
A system on module needs a carrier-board contract, not a connector map
An AMD Xilinx system on module can speed up a schedule, but only if the carrier-board boundary is designed as an engineering contract. Five steps for a repeatable integration.
Zero-copy in an Embedded Linux camera pipeline means moving ownership, not pixels
Zero-copy does not mean no data movement. It means avoiding unnecessary CPU-mediated copies between pipeline stages, and it is harder than it sounds.
A watchdog is only useful when feeding it proves the system is healthy
A watchdog fed from a periodic loop proves very little. How to gate each feed on real health signals and recover predictably from an unrecoverable state.
A secure OTA update is a controlled recovery process, not a file download
Five questions a dependable embedded OTA design should answer: authenticity, integrity, compatibility, recovery and observability.
Embedded Linux board bring-up starts long before the first boot log appears
The fastest teams reduce uncertainty layer by layer. A five-step approach to bringing up a new Embedded Linux board without debugging everything at once.