**6. Conclusions**

While data collected by IoT smart city applications are a central asset in supporting management and planning decisions for many communities, designing and deploying IoT solutions is still challenging due to system integration complexity, reliability limitations, and cost. We presented a cloud data storage and visualization system for smart cities, leveraging reliable existing technology to integrate a complete IoT monitoring solution hosted in AWS and costing under USD 26/year for long-term data storage and USD 0.0204/hour of use for MySQL database and Grafana servers. By using this cloud-based solution together with TTN infrastructure and commercial LoRaWAN sensors, users can collect, store, and visualize datasets to address their needs and integrate their own services. We demonstrated the use of the system for a flood warning system application example with river and weather LoRaWAN sensors. The cloud-based system design uses serverless data ingestion to provide a simple and cost-effective data storage solution that is independent of other services such as data visualization. An on-demand database and visualization servers offer flexibility to adapt to application needs while saving costs and simplifying maintenance operations. Furthermore, we explored the different AWS tiers and their respective reliability/cost tradeoff so users can make informed decisions when tailoring our system to their own application. As opposed to focusing mainly on the example application, as commonly seen in the literature, we highlight common tasks that are required by an IoT project and share our insights in leveraging modern cloud services to simplify IoT backend system design and optimize costs.

As a future research avenue, we intend to explore the use of new serverless cloud backend architectures in smart city IoT applications and investigate practical tradeoffs to server solutions. We intend to analyze in particular the on-demand allocation of computational resources as we scale the number of sensors, total sensor data rate, and number of clients connecting to user interfaces in cloud IoT backend systems. We also intend to explore the integration of modeling and simulation tools with IoT data acquisition systems while efficiently allocating computational resources.

**Author Contributions:** Conceptualization, V.A.L.S., B.C. and J.L.G.; methodology, V.A.L.S., L.A., A.M., G.M., K.T. and C.T.; software, V.A.L.S., L.A., A.M., G.M., K.T. and C.T.; validation, V.A.L.S.; data curation, V.A.L.S.; writing—original draft preparation, L.A., A.M., G.M., K.T. and C.T.; writing—review and editing, V.A.L.S., J.N., B.C. and J.L.G.; supervision, V.A.L.S., B.C. and J.L.G.; funding acquisition, B.C. and J.L.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** We acknowledge support from the US National Science Foundation through the award number 1735587.

**Data Availability Statement:** Code and instructions to setup the cloud infrastructure described in this paper are available at https://github.com/uva-hydroinformatics/iot-cloud-platform (accessed on 14 April 2023).

**Acknowledgments:** The authors would like to recognize Ruchir Shah for his valuable comments on and assistance with this work.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
