Open Access

ARTICLE

Cost and Time Optimization of Cloud Services in Arduino-Based Internet of Things Systems for Energy Applications

Reza Nadimi^1,*, Maryam Hashemi², Koji Tokimatsu³

1 Department of Innovation Science, School of Environment and Society, Institute of Science Tokyo, Minato-ku, Tokyo, 108-0023, Japan
2 Computer Engineering Department, Azad University, Saveh City, 81318, Iran
3 Department of Transdisciplinary Science and Engineering, School of Environment and Society, Institute of Science Tokyo, Yokohama, 226-8503, Japan

* Corresponding Author: Reza Nadimi. Email:

Journal on Internet of Things 2025, 7, 49-69. https://doi.org/10.32604/jiot.2025.070822

Received 24 July 2025; Accepted 29 August 2025; Issue published 30 September 2025

Abstract

Existing Internet of Things (IoT) systems that rely on Amazon Web Services (AWS) often encounter inefficiencies in data retrieval and high operational costs, especially when using DynamoDB for large-scale sensor data. These limitations hinder the scalability and responsiveness of applications such as remote energy monitoring systems. This research focuses on designing and developing an Arduino-based IoT system aimed at optimizing data transmission costs by concentrating on these services. The proposed method employs AWS Lambda functions with Amazon Relational Database Service (RDS) to facilitate the transmission of data collected from temperature and humidity sensors to the RDS database. In contrast, the conventional method utilizes Amazon DynamoDB for storing the same sensor data. Data were collected from 01 April 2022, to 26 August 2022, in Tokyo, Japan, focusing on temperature and relative humidity with a resolution of one minute. The efficiency of the two methods—conventional and proposed—was assessed in terms of both time and cost metrics, with a particular focus on data retrieval. The conventional method exhibited linear time complexity, leading to longer data retrieval times as the dataset grew, mainly due to DynamoDB’s pagination requirements and the parsing of payload data during the reading process. In contrast, the proposed method significantly reduced retrieval times for larger datasets by parsing payload data before writing it to the RDS database. Cost analysis revealed a savings of $1.56 per month with the adoption of the proposed approach for a 20-gigabyte database.

---

Keywords

---