20/05/2026
5 tender criteria for LoRaWAN at scale.
For tenders covering more than ~1,000 connected luminaires, network behaviour in the first minutes after a coordinated power-on is a measurable acceptance criterion — and one a vendor can meaningfully be held to. A short list we apply when reviewing or drafting these tender sections:
1. Acceptance window. State a measurable post-power-on commissioning target — e.g. "≥ 95 % of devices joined within 30 minutes for a 5,000-node region." Anchor it to the ISM duty-cycle limit at the relevant SF, not to a vendor claim.
2. Behaviour under join contention. Require a documented jitter range on first-boot join attempts plus a published retry-backoff policy. "Joins on boot" without further specification is not acceptable.
3. Session-state persistence across power loss. The device must not perform a fresh OTAA join as the default response to a brown-out or scheduled power cycle. ABP fallback alone is not sufficient.
4. Rejoin trigger policy. Rejoin should be driven by signal/link conditions (link loss, ADR drift, ack-fail thresholds), not by fixed timers. Vendor must publish the trigger logic.
5. Population-level boot pattern. The firmware must support a non-uniform boot sequence (randomised window, group staggering) configurable to the size and gateway density of the target deployment.
Why these five: they are the criteria most likely to be missed in a tender draft and most likely to surface only after rollout, when remediation is most expensive.
→ Want a 1-pager with these criteria written in tender-ready language? Comment "spec" or DM us.
