Performance Evaluation and Scalability Analysis of FIWARE, ThingsBoard, and EdgeX Foundry IoT Platforms

JSYS 2025 October Papers Submission5 Authors

Keywords: IoT platforms, performance evaluation, scalability analysis, FIWARE, ThingsBoard, EdgeX Foundry, load testing

Abstract:

This paper presents a comprehensive performance evaluation and scalability analysis of three prominent opensource IoT platforms: FIWARE, ThingsBoard, and EdgeX Foundry. Using systematic load testing with JMeter at three distinct scales (10×100, 100×1000, 1000×1000), we evaluate platform performance across
single-node and multi-node deployments. Our experimental results demonstrate significant scalability differences: FIWARE achieves 20-25% system CPU utilization in multi-node deployments compared to 37-50% in single-node configurations but experiences complete failure under maximum load; ThingsBoard shows counterintuitive scaling behavior with multi-node deployments consuming 105% CPU versus 75% in single-node setups; while EdgeX Foundry maintains exceptional stability with error rates below 1% and optimal resource utilization across all tested scenarios.
Key findings reveal measurable performance improvements in multi-node FIWARE deployments which reduce system CPU usage from 37-50% to 20-25% under low load while maintaining zero error rates. ThingsBoard demonstrates limited horizontal scaling benefits with error rates improving marginally from 8% to 5% at medium load but shows concerning resource utilization patterns. EdgeX Foundry achieves superior cache performance with >90% Redis hit rates and sustained throughput while maintaining stability even under extreme 1000×1000 load conditions. Database performance analysis shows FIWARE MongoDB replica sets exhibit suboptimal behavior with I/O operations increasing from 30 to 190 ops/sec in multi-node configurations, while EdgeX Foundry’s Redis implementation maintains consistent high performance. These quantitative results provide critical insights for IoT platform selection and deployment strategies in large-scale environments.

Area: Configuration Management for Systems

Type: Problem

Revision: No

Submission Number: 5