Source-linked AI summary
Slotted ALOHA on LoRaWAN - Design, Analysis, and Deployment
Tommaso Polonelli, Davide Brunelli, Achille Marzocchi, Luca Benini
TL;DR
LoRaWAN scalability is challenged by massive device populations, long packet airtimes, and synchronization limitations. The paper overlays synchronized Slotted-ALOHA on standard LoRaWAN and reports improved throughput and fewer collisions, while identifying synchronization accuracy and multi-channel evaluation as boundaries.
Problem
LoRaWAN scalability is challenged by large device populations, long packet airtimes, and Pure-ALOHA access inefficiency.
Method
The paper models LoRaWAN traffic and overlays Slotted-ALOHA using synchronization, application-layer random backoff, and unmodified LoRaWAN firmware.
Results
Measured throughput reaches 15% for S-LoRaWAN versus 2.6% for standard LoRaWAN, a 5.8x improvement, with 26% fewer packet collisions in a 24-node deployment.
Takeaways & Limitations
S-LoRaWAN provides a deployable throughput improvement while preserving Class A low-power capabilities and requiring only 8 additional downlink bytes.
Takeaways & Limitations
Synchronization accuracy is constrained by gateway process management, with average error of 5.37ms and maximum error of 37.08ms; multi-channel performance remains outside the scope.
Abstract
from arXiv · showhide
LoRaWAN is one of the most promising standards for long-range sensing applications. However, the high number of end devices expected in at-scale deployment, combined with the absence of an effective synchronization scheme, challenge the scalability of this standard. In this paper, we present an approach to increase network throughput through a Slotted-ALOHA overlay on LoRaWAN networks. To increase the single channel capacity, we propose to regulate the communication of LoRaWAN networks using a Slotted-ALOHA variant on the top of the Pure-ALOHA approach used by the standard; thus, no modification in pre-existing libraries is necessary. Our method is based on an innovative synchronization service that is suitable for low-cost wireless sensor nodes. We modelled the LoRaWAN channel with extensive measurement on hardware platforms, and we quantified the impact of tuning parameters on physical and medium access control layers, as well as the packet collision rate. Results show that Slotted-ALOHA supported by our synchronization service significantly improves the performance of traditional LoRaWAN networks regarding packet loss rate and network throughput.
1. Introduction
The paper examines LoRaWAN scalability and channel throughput for large deployments, then proposes a synchronized Slotted-ALOHA overlay to improve performance without changing standard firmware.
- Motivation: LoRaWAN scalability remains insufficiently characterized for large networks with long packet airtimes.Prior results cited in the paper report only 120 nodes under static configurations and up to 1600 with dynamic configurations.
- Motivation: Existing scalability studies often omit downstream traffic, interference, clock drift, and gateway or server delays.
- Approach: The study analyzes channel throughput and scalability using a simulator for LoRaWAN Class A under varying device counts and airtimes.
- Approach: A synchronized Slotted-ALOHA overlay reduces protocol-overhead impact and improves throughput over Pure-ALOHA.The synchronization service achieves 5.37ms accuracy and over 99% success, while tests with 24 nodes show 2x throughput improvement and 5.8x improvement in crowded conditions.
- Paper structure: The paper proceeds from channel modeling and simulation to synchronization, implementation, throughput evaluation, synchronization-error analysis, and conclusions.
2. Related Works
Related work identifies Pure-ALOHA access as a major source of LoRaWAN inefficiency and motivates Slotted-ALOHA, whose slot structure separates free, collided, and successful transmissions.
- LPWAN context: Large-scale IoT deployments require low-power, long-range communication across applications such as smart cities and transportation.The paper notes forecasts of 21B connected devices by the end of the decade and an annual growth rate of 32%.
- Alternative protocols: DASH7 offers low latency, mobility support, and multi-year battery life, but uses more power per bit than LoRaWAN.
- LoRaWAN context: LoRaWAN architecture connects sensor nodes to a gateway and defines Class A, B, and C end-device operation.
- Access limitations: ALOHA-style access allows transmissions without carrier sensing, causing inefficiency and rapid saturation as traffic increases.One cited simulation reaches a maximum expected channel-capacity usage of 18% at a link load of 0.48.
- Slotted-ALOHA basis: Slotted ALOHA divides channel time into fixed slots, allowing one transmission to succeed while simultaneous transmissions collide and unused slots remain free.
3. LoRa physical layer and LoRaWAN
The paper describes LoRa modulation, regional channel allocation, orthogonal spreading factors, regulatory constraints, and LoRaWAN class-specific transmission behavior.
- LoRa physical layer: LoRa uses chirp spread spectrum modulation whose frequency changes over time across a defined bandwidth.The modulation supports long-range communication and can tolerate frequency differences between gateways and transmitters.
- LoRa physical layer: Code rate, spreading factor, and bandwidth are the principal parameters affecting LoRa modulation.
- Channel configuration: In the Italian European deployment studied, LoRaWAN uses channels in the 863–870MHz ISM band, with 125KHz channels supporting 0.3kbps to 5kbps data rates.Channel ten is allocated to FSK modulation with 250KHz bandwidth.
- Channel configuration: Different spreading factors are orthogonal, allowing a gateway to receive multiple transmissions simultaneously on the same channel.
- Regulatory constraints: Regional ISM rules limit each ALOHA device to occupying no more than 1% of channel time, while transmissions across permitted sub-bands can increase overall throughput.
- LoRaWAN operation: Class A schedules uplinks at the end node and permits downlinks only in receive windows opened after an uplink.The paper's experiments use the same data rate and channel for uplink and downlink and disable RX2 for static evaluation.
4. LoRaWAN MAC Model
LoRaWAN Class A uses a Pure-ALOHA-like MAC whose half-duplex overhead substantially constrains single-channel throughput. The model accounts for uplink, downlink, RX1 timing, and packet airtime, then evaluates these effects through simulation.
- MAC operation: Class A LoRaWAN is based on Pure-ALOHA, with downlink communication constrained to two short reception intervals after each uplink.This battery-oriented design permits bidirectional communication while limiting when downlinks can occur.
- Pure-ALOHA model: A 2T vulnerability period determines successful Pure-ALOHA transmission when the channel remains free throughout that interval.The model represents aggregate packet generation as a Poisson source and includes retransmitted packets in total traffic.
- LoRaWAN timing: With SF 12, BW 125 KHz, CR 4, and 25-byte payloads, airtime is 1253ms while the corresponding ACK lasts 530ms.A one-second RX1 delay further shapes the opportunity and likelihood of successful transmissions.
- Throughput model: LoRaWAN overhead expands the vulnerability period from 2T to 2·(2.22T), reducing effective throughput to a maximum of 8% at G equal to 0.25.The 2.22T slot-time representation includes the uplink, downlink, and RX1-delay overhead.
- Simulation: The simulator evaluates unconfirmed and confirmed single-channel links across packet-airtime configurations, with shorter-than-one-second airtime slightly improving performance.When packet airtime exceeds the RX1 delay, the confirmed-link collision condition is especially restrictive.
- Slotted operation: The proposed Slotted-ALOHA arrangement fits the LoRaWAN link components into one slot and reaches 16% maximum channel throughput.This is twice the classic half-duplex LoRaWAN value reported in the model.
5. Real Time Clock Synchronization over LoRaWAN stack
The paper adds lightweight RTC synchronization over LoRaWAN so end devices can align S-ALOHA slots without significant communication overhead. Gateway timestamp processing and crystal drift limit synchronization accuracy and determine deployment timing margins.
- Synchronization design: Clock synchronization aligns end devices for S-ALOHA slots, using a lightweight method tailored to LoRaWAN nodes.The method uses an RTC with an inexpensive 32.768KHz crystal and periodically redistributes the gateway’s time reference.
- Synchronization design: Each node updates its RTC after an uplink by receiving a gateway timestamp in the first available acknowledgment window.The algorithm records transmission-end timestamps at the gateway and node, then returns the gateway timestamp through the scheduled RX window.
- Synchronization design: The node calculates a timing offset because acknowledgment delay makes the received gateway timestamp no longer current.The calculation uses the interval between the node’s transmission end and the response packet; propagation time is negligible relative to these delays.
- Accuracy and deployment limits: Increasing synchronization-requesting devices reduces accuracy, with an average error of 5.37ms and a maximum of 37.08ms.The uncertainty is attributed partly to gateway operating-system process management and must be considered to avoid adjacent-slot overlaps.
- Accuracy and deployment limits: An 80ppm crystal can generate 200ms of error every 40 minutes, guiding the slot period and RTC refresh rate.The required timing margin depends on application and transmission period.
6. S-ALOHA implementation over LoRaWAN
S-LoRaWAN implements slotted access above the LoRaWAN MAC through application-layer scheduling, acknowledgments, and randomized retransmission backoff. This preserves compatibility while avoiding random channel selection for initial transmissions.
- Slotted transmission: End devices schedule uplinks only at the beginning of a slot when their RTC triggers, rather than selecting transmission timing randomly.The application-layer schedule avoids evaluating random channel access across multiple channels.
- Retransmission control: Lost packets trigger retransmission in a randomly selected slot using a configurable multiple of the single-slot time.The random backoff reduces repeated interference corner cases among devices.
- Implementation: S-LoRaWAN is implemented at the application layer with two timers and LoRa transmission libraries.Its state machine includes timers for first-slot transmission, random retransmission backoff, and acknowledgment handling.
- Retransmission control: The backoff state ends when TimerBackOff expires, after which the device enters Send and retransmits the message.This is the state-machine transition used for delayed packet recovery.
7. S-ALOHA evaluation in a real deployment
The paper evaluates S-LoRaWAN on custom sensor hardware and an indoor CINECA deployment using packet counters and controlled traffic tests. The implementation doubles theoretical throughput, while deployment results show substantially higher measured throughput than standard LoRaWAN.
- Hardware and deployment: The custom node combines an STM32L476, RFM95W LoRa transceiver, energy harvester, temperature-humidity sensor, and expansion connector.Its sleep current is 4µA at 3V with the RTC enabled, while the MCU draws 8.25mA in 48MHz run mode.
- Hardware and deployment: Indoor tests at CINECA used distributed sensor nodes to monitor temperature, demonstrating a real deployment setting for S-LoRaWAN.The experiments used an industrial IoT gateway with LoRa radio modules and Node-RED application software.
- Results: S-ALOHA enables doubling exchanged data or increasing the overall number of sensors in the tested network.This conclusion is tied to the CINECA deployment results.
- Evaluation method: Performance evaluation recorded per-device sent and lost packet counters through a database managed by the LoRaWAN server.The counters were integrated into the S-ALOHA state machine to compare network behavior.
- Evaluation method: The tests varied device count and message frequency while holding payload size and spreading factor constant, using confirmed messages and application-layer backoff.Results compare offered and successfully disposed traffic after temporal normalization.
- Results: The maximum theoretical throughput doubled and followed a trend similar to the standard S-ALOHA statistical model.Throughput was computed from channel traffic and the average percentage of successfully received packets.
8. Comparison between LoRaWAN and Slotted LoRaWAN
Using identical radio configurations and injected traffic, the paper compares standard LoRaWAN with S-LoRaWAN under payload-time traffic accounting. S-LoRaWAN achieves a 5.8x measured throughput improvement in the reported test.
- Comparison setup: The comparison uses the same radio parameters and injected traffic for standard LoRaWAN and S-LoRaWAN.Traffic and throughput are calculated using payload time, with slot tolerance included in protocol overhead.
- Measured comparison: 15% versus 2.6% measured throughput yields a 5.8x improvement for S-LoRaWAN over standard LoRaWAN.The reported values correspond to S-LoRaWAN and standard LoRaWAN, respectively.
- Measured comparison: Standard LoRaWAN produces more overlapping frames, leading to greater backoff use and fewer average transmitted packets.In the reported test, the Pure-ALOHA channel is heavily crowded and has a 7% transmission success rate.
- Interpretation: The measured improvement exceeds the theoretical 2x S-ALOHA versus P-ALOHA difference because the protocols reach maximum throughput at different traffic loads.The comparison therefore depends on the tested operating point, not only on the theoretical maximum.
9. Synchronization analysis in Test 3
In Test 3, synchronization achieved a 99.6% slot success rate across 24 nodes, while a 400 ms tolerance was exceeded five times. Reducing the tolerance interval can preserve acceptable success while increasing available slots and throughput.
- Synchronization performance: 99.6% slot success was achieved across 24 nodes, with five failed slots overall.The failures included overlaps between adjacent slots.
- Synchronization performance: Five tolerance-interval exceedances occurred at the 400 ms limit.Two overflows were generated by an end-node reboot caused by firmware issues.
- Tolerance-interval trade-off: A 100 ms tolerance produced an error rate practically equal to Test 3, while 50 ms remained acceptable.The tolerance interval was evaluated as a way to reduce overhead without sacrificing acceptable slot success.
- Tolerance-interval trade-off: Shortening the tolerance interval reduces protocol overhead and allows more slots per unit time.The relationship between tolerance-interval length and failed slots is presented in Figure 14(b) and Table 4.
10. Conclusions
The paper presents S-LoRaWAN as a synchronized Slotted-ALOHA protocol for low-cost, low-power IoT devices. It reports higher throughput and fewer collisions without modifying the existing LoRaWAN software stack.
- Conclusions: S-LoRaWAN provides a reliable synchronization method for a Slotted-ALOHA protocol suitable for low-cost and low-power IoT devices.The paper frames the method as an implementation for LoRaWAN networks.
- Conclusions: 2x theoretical network throughput improvement and up to 5.8x measured improvement were reported.The measured improvement was obtained in a high-traffic setup.
- Conclusions: 26% fewer packet collisions were demonstrated in a real-life deployment with 24 nodes operating for hours.This result is reported alongside the measured throughput improvement.
- Conclusions: The proposed S-LoRaWAN adds only 8 downlink bytes with negligible power-consumption consequences.The paper identifies low power consumption as important for battery-operated devices.
- Conclusions: S-LoRaWAN requires no changes to the LoRaWAN software stack and can run on unmodified firmware.This supports deployment as an overlay on existing LoRaWAN implementations.
Appendix A
The appendix models LoRa transmission duration from preamble and physical-message components. It defines the symbol and message parameters used to calculate packet timing and transmission data rate.
- Transmission-time model: LoRa transmission duration is the sum of preamble time and physical-message time.The duration depends on Spreading Factor, Bandwidth, and Coding Rate.
- Transmission-time model: The preamble and physical-message terms are calculated from their transmitted symbol counts.N_preamble counts preamble symbols, while N_PHY counts physical-message symbols.
- Physical-message calculation: Symbol duration T_sym is measured in seconds and depends on Spreading Factor and Bandwidth.T_sym is used in calculating the transmission-time terms.
- Physical-message calculation: N_PHY depends on payload length, CRC presence, clock-drift protection, and modulation parameters.The passage defines PL, CRC, and DE, including DE values for SF12 and SF11.