Evaluation of Performance Gain Using Parallelism in IoT Data from Smart Cities

• Matheus Tregnago UCS
• Ricardo Bregalda UCS
• Samuel Francisco Ferrigo UCS

DOI: https://doi.org/10.5753/eradrs.2025.6527

Abstract

This paper evaluates the use of different parallelism techniques for processing IoT data in smart cities. For this, the parallelism techniques OpenMP, MPI, Multiprocessing, and Dask were used to identify sensor failures and assess the performance gain, achieving a speedup of up to 5.16x with Dask. The results highlight the advantages of using parallelism in processing large volumes of data.

Keywords: Parallel and Distributed Algorithms

References

Bekkai, B., Bendjenna, H., and Kitouni, I. (2021). Internet of things: A recent survey. In 2021 International Conference on Recent Advances in Mathematics and Informatics (ICRAMI), page 1–9, Tebessa, Algeria. IEEE.

Carleton (2024). Introduction to openmp. Disponível em: [link]. Acesso em: 24 nov. 2024.

Carvalho, M. D. S. d., Ferrigo, S. F., Dal Bó, G., Fachinelli, A. C., Perini, R. d. L., and Mosche, S. d. A. (2024). Iot para cidades inteligentes: habilitação tecnológica em uma cidade da serra gaúcha. Disponível em: [link]. Acesso em: 20.jan.2025.

Dask (2024). Dask: Scalable parallel computing in python. Disponível em: [link]. Acesso em: 24 nov. 2024.

MPI (2024). Mpi 4.1: The complete mpi standard. Disponível em: [link]. Acesso em: 24 nov. 2024.

Python (2024). multiprocessing — process-based parallelism. Disponível em: [link]. Acesso em: 24 nov. 2024.