IoT Device Management

When to use

• Designing zero-touch provisioning flows for manufacturing at scale
• Implementing OTA firmware updates with A/B partitions and rollback
• Building device twin / shadow state synchronization between device and backend
• Setting up command-and-control (request-response, fire-and-forget, direct methods)
• Managing fleet operations: group updates, health dashboards, decommissioning

Core principles

1. Provision from hardware roots — private keys belong in secure elements or TPMs; never in firmware flash or environment variables
2. A/B partitions over in-place updates — the inactive slot absorbs the risk; rollback is a reboot, not a crisis
3. Twin delta is the source of truth — desired vs reported gap tells you exactly what the fleet hasn't converged on yet
4. Commands must be idempotent and TTL-bound — offline devices wake up; stale reboot commands should not execute
5. Fleet operations are canary-first — 1% → 10% → 50% → 100% with automated pause on error rate threshold

Reference Files

• references/provisioning.md — zero-touch provisioning flow, certificate-based onboarding, DPS allocation policies, device registry structure
• references/ota-updates.md — A/B partition strategy, delta updates with bsdiff, rollback mechanisms, staged campaign management
• references/device-twin-and-fleet.md — twin/shadow state pattern, synchronization, conflict resolution, command patterns, fleet operations