Development of an expandable network of IoT (Internet of Things) devices for data acquisition and processing in real time
In this thesis, we worked on a small scale sensor network application, experimenting with different combinations of software and hardware platforms, communication protocols and data storage solutions. We focused on alert-based sensor systems that are able to efficiently endure through long time span...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Language: | en_US |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/11610/19573 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In this thesis, we worked on a small scale sensor network application, experimenting with different combinations of software and hardware platforms, communication protocols and data storage solutions. We focused on alert-based sensor systems that are able to efficiently endure through long time spans consisting of resource constrained devices. Our study exposes the major challenges that arise when dealing with this kind of systems concerning hardware limitations, network throughput, data processing, secure transmissions and long term consistency. As a test bench, we used low power consumption devices with networking capabilities supporting numerous communication protocols. To “glue” our nodes together, we used the Message Queuing Telemetry Transport (MQTT) [Appx.7] (i.e., Appendix entry #7), [1] protocol conjoined with Transport Layer Security (TLS) [Appx.13] to harden security across the communication channels. By using MQTT, we also focused on the formalization of this machine-to-machine (M2M) [Appx.6] protocol’s core elements, by defining a solid message/topic structure throughout our network. Various sensors were used, gathering data on different platforms and multiple network layers, in order to analyze the behavior of the network from core to Edge [Appx.2]. We combined our locally acquired data with input from popular cloud based services and experimented with data exchange formats and database storage technologies. Finally, some benchmarking was performed on every node, in order to evaluate our system’s processing throughput and optimize our settings to enhance robustness and efficiency. We managed to build a real time monitoring system able to endure in realistic conditions in our university campus. The development of this thesis led to the paper entitled: “Towards standardization of MQTT-alert-based sensor networks: protocol structures formalization and low-end node security” [67], that was presented in the 13th International Symposium of Industrial Embedded Systems (SIES 2018), in June 2018. |
|---|