TO: Prof. Ellis
FROM: Kiara Candelario
DATE: 3/03/2021
SUBJECT: 500-Word Summary of Article About Comparing Non-Relational and Relational Databases.
The following is a 500-word summary of a peer-reviewed article about testing and comparing MongoDB and MySQL using IoT data on a virtual machine. The Internet of things is a system that consists of sensing, and collecting data, and it’s becoming a large aspect in many industries. According to the author, ” using IoT technology generates a large amount of heterogeneous data like texts, numbers, audio, videos, and pictures. These types of data need to be transferred, processed and stored” (Eyada et al., 2020, p. 110656). IoT data comes from different sources, and a database management system can assist with storing the amount of data that IoT creates. Relational DBMS’s use SQL,which is a popular system, but IoT data is heterogeneous, and it can negatively affect the database’s performance. NoSQL database, also known as a non-relational database, is the best option for IoT data due to storing unstructured data and is schema-free. NoSQL also has high scalability and availability. Cloud computing can deal with large amounts of data, and databases use cloud computing to improve consistency, availability, and tolerance.
MySQL is a relational database system that uses SQL to store data in tables and needs a pre-defined schema. Any change to the schema can hinder the performance and takes the database offline. MongoDB is a non-relational database system that is document-oriented, and it stores data as BSON objects. It has quick query access, and a structure does not need to be declared. MongoDB has different features that provide better performance based on long-term storage of large amounts of data and flexibility to work. The current experiment will solve the previous limitations that the other experiments had by enhancing both databases and not limiting the number of sensor nodes.
MongoDB and MySQL will store the IoT information, and it is base on the data collected from air pollution indoors and outdoors. In the MySQL database setup, two tables are created named station_location and town_name, which manage the station’s location and the sensor nodes. In the MongoDB Database Setup, two collections are made, where the first collection saves every station’s location. The second collection is the sensor table for all the sensors in the station. Node.JS is the server language that is used to process the collected data. Ubuntu 16. 04 LTS is the operating system installed on the virtual machine to setup MongoDB, MySQL, and Node.JS. Amazon Web Service’s Elastic Compute Cloud is the virtual machine that is used to establish the environment.
The experiment was conducted based on increasing the workload of each database latency, database size, and the number of sensor nodes. The impact of increasing the workload resulted in a latency decrease in the MongoDB database compared to the MySQL database. The impact of increasing the workload on database sizes demonstrates that MySQL outperforms MongoDB. Lastly, increasing the number of sensor nodes that connect to each station resulted in MongoDB outperforming MySQL significantly. The results demonstrate that MongoDB outperforms MySQL due to MySQL performance loss when increasing the workload.
Reference:
M. M. Eyada, W. Saber, M. M. El Genidy and F. Amer, “Performance Evaluation of IoT Data Management Using MongoDB Versus MySQL Databases in Different Cloud Environments,” in IEEE Access, vol. 8, pp. 110656-110668, 2020, doi: 10.1109/ACCESS.2020.3002164.
