LoRaConnect: Unlocking HTTP Potential on LoRa Backbones for Remote Areas and Ad-Hoc Networks

TL;DR

LoRaConnect enables HTTP access over LoRa networks by implementing caching, message slicing, and synchronization, allowing web applications to function in remote and ad-hoc areas despite LoRa's limitations.

Contribution

The paper introduces LoRaConnect, a system that facilitates HTTP communication over LoRa networks through novel hardware and software mechanisms, addressing payload and collision challenges.

Findings

Achieves 1.18 KB/s throughput for HTTP over LoRa

Reduces webpage access delay to around 1.3 seconds in baseline

Demonstrates multi-fold performance improvement over benchmarks

Abstract

Minimal infrastructure requirements make LoRa suitable for service delivery in remote areas. Additionally, web applications have become a de-facto standard for modern service delivery. However, Long Range (LoRa) fails to enable HTTP access due to its limited bandwidth, payload size limitations, and high collisions in multi-user setups. We propose LoRaConnect to enable HTTP access over LoRa. The LoRaWeb hardware tethers a WiFi hotspot to which client devices connect and access HTTP resources over LoRa backhaul. It implements caching and synchronization mechanisms to address LoRa's aforementioned limitations. It also implements a message-slicing method in the application layer to overcome LoRa's payload limitations. We evaluate the proposed system using actual hardware in three experimental setups to assess the baseline performance, ideal scenario, and practical application scenario with Frequency Hopping Spread Spectrum (FHSS). Additionally, it implements a ping operation to demonstrate Internet capability and extensible nature. LoRaWeb achieves an average throughput of 1.18 KB/S approximately, with an access delay of only 1.3 S approximately for a 1.5KB webpage in the baseline setup. Moreover, it achieves an access delay of approximately 6.7 S for a 10KB webpage in the ideal case and an average end-to-end delay of only 612 ms approximately in the FHSS-based setup. Comparison with benchmark suggests multi-fold improvement.