IoT Connectivity Trade-Offs: Range, Bandwidth, Power, Cost

Every connectivity technology decision in an IoT design is really a four-way trade against physics: you cannot simultaneously maximize range, bandwidth, and battery life while minimizing cost. Choosing Wi-Fi over LoRaWAN, or NB-IoT over BLE, fixes not just a radio but a whole set of downstream constraints — antenna design, battery size, cloud ingestion pattern, and unit cost at volume. Getting this wrong at the architecture stage is expensive to fix later, because it usually means a board respin or a field truck-roll to swap radios.

The Four Axes

• Range: the maximum reliable link distance, governed by transmit power, receiver sensitivity, path loss, and obstructions. Quoted "line of sight" ranges are typically 3–10× optimistic vs. real indoor/urban deployments.
• Bandwidth (data rate): how many bits/second the link can carry, which determines payload size, update frequency, and whether firmware-over-the-air (FOTA) images are practical to push.
• Power: energy per bit transmitted, plus idle/listening current. This drives battery size, replacement interval, and whether energy harvesting is viable.
• Cost: silicon/module BOM cost, certification cost (FCC/CE/regional), and recurring cost (cellular data plans, gateway infrastructure, licensed spectrum fees).

These axes are coupled through basic radio physics, not just market positioning. Two relationships matter most:

1. Range vs. power — the link budget. To reach further without more power, you either lower the data rate (more energy per bit, e.g., LoRa's spreading factor) or increase antenna gain/receiver sensitivity.
2. Bandwidth vs. power — Shannon's limit implies higher throughput needs more SNR, which usually means more transmit power or shorter range for the same power.

Link budget (dB) = Tx power + Tx antenna gain − path loss
                    + Rx antenna gain − Rx sensitivity − margin

A link closes when the budget is ≥ 0 dB. Every dB of range you want costs power, receiver complexity, or lower data rate — there is no free lunch.

Path Loss Sets the Baseline

Free-space path loss: FSPL(dB) = 20·log₁₀(d) + 20·log₁₀(f) + 32.44 (d in km, f in MHz).

Doubling distance costs 6 dB; doubling frequency also costs 6 dB. This is why sub-GHz LPWAN (868/915 MHz) reaches farther than 2.4 GHz Wi-Fi/BLE for the same power — not because the protocol is "better," but because path loss is lower at lower frequency. Indoor environments add 10–30 dB of extra loss from walls and floors, which is why rated ranges collapse indoors.

Comparing the Common Choices

Numbers are order-of-magnitude, real-world figures vary heavily with duty cycle, module vendor, and RF environment; always validate against the specific module's datasheet link budget.

Worked Example: Battery Life for a Sensor Node

Compare a BLE sensor node transmitting a 20-byte payload once per minute against an NB-IoT node sending the same payload once per hour, both on a 1000 mAh cell, both otherwise idle at 3 µA sleep current.

BLE node:

• Tx event: ~8 mA for ~5 ms (connection event with a few packets) → charge per event ≈ 8 mA × 0.005 s = 40 µA·s = 0.0111 µAh
• Events/day: 1440 (once/minute) → daily Tx charge = 1440 × 0.0111 µAh ≈ 16 µAh
• Sleep charge: 3 µA × 24 h = 72 µAh/day
• Total ≈ 88 µAh/day → 1000 mAh / 0.088 mAh ≈ 11,360 days ≈ 31 years (dominated by sleep current and self-discharge in practice, but shows Tx is nearly free at this rate)

NB-IoT node:

• Tx event: ~200 mA for ~2 s (attach + send small payload, PSM-optimized) → 200 mA × 2 s = 400 mA·s = 0.111 mAh per event
• Events/day: 24 (once/hour) → daily Tx charge = 24 × 0.111 mAh ≈ 2.67 mAh
• Sleep charge (PSM idle, more leakage than BLE, assume 10 µA): 10 µA × 24 h = 240 µAh = 0.24 mAh/day
• Total ≈ 2.91 mAh/day → 1000 mAh / 2.91 mAh ≈ 344 days ≈ 11 months

Check: the ratio of Tx energy per event (200 mA·2 s = 400 mA·s vs. 8 mA·5 ms = 0.04 mA·s) is ~10,000×, while NB-IoT sends 60× less often, so the net Tx-energy-per-day ratio is ~10,000/60 ≈ 167×. This is confirmed directly by the daily Tx charges above: 2.67 mAh vs. 0.016 mAh ≈ 167×. This matches the intuition that cellular's long attach/registration overhead and higher link budget dominate power budget, while BLE's short-range, low-power radio wins decisively when a gateway is nearby. The trade is direct: NB-IoT buys kilometers of range and no local infrastructure, at a two-orders-of-magnitude power cost.

Practical Selection Framework

• If nodes are within 10–50 m of a hub you control: BLE or Zigbee/Thread — lowest power, lowest cost, no recurring fees, but requires a local gateway.
• If nodes are spread over km² with no existing infrastructure and low data rates suffice: LoRaWAN (private gateway, no carrier fee) or NB-IoT/LTE-M (carrier coverage, recurring SIM cost) — trade power and cost for range.
• If payloads are large or need low latency (video, frequent firmware updates): Wi-Fi or LTE-M/5G RedCap — trade power and cost for bandwidth.
• If the device is mobile across wide areas (vehicles, asset tracking): cellular (NB-IoT/LTE-M/5G) — the only option combining range and mobility without dense infrastructure.
• If unit cost at high volume is the primary constraint: BLE/Zigbee silicon is typically the cheapest; cellular modules carry both higher BOM and recurring plan costs; factor certification cost (cellular carrier certification adds significant NRE vs. unlicensed-band radios).

Key Takeaways

• Range, bandwidth, power, and cost are coupled through the link budget and Shannon capacity — improving one typically degrades at least one other.
• Lower frequency and lower data rate buy range for the same transmit power (LoRa's spreading factor is a direct example); this is physics, not marketing.
• Short-range protocols (BLE, Zigbee, Thread, Wi-Fi) minimize power and cost per node but require gateway infrastructure within range.
• LPWAN (LoRaWAN, NB-IoT, LTE-M) trades higher per-transmission energy and/or recurring carrier cost for multi-kilometer range without dense local infrastructure.
• Always validate vendor "typical range" figures against your own link budget and real-world obstruction losses — rated ranges assume line-of-sight conditions rarely seen indoors or in urban deployments.
• Battery-life estimates must account for attach/registration overhead (especially cellular), not just raw transmit current, since overhead often dominates total energy per message.