Authors : Minh-Son Dao, Yen-Nhi Tran-Thi,, Van-Cuong Mai and Tuan-Anh Nguyen-Gia
Abstract: The era of big data coming from everyone, everything and everywhere raises a big challenge to data science, especially, the database category. Meanwhile, SQL databases guarantee ACID (Atomicity, Consistency, Isolation, Durability) and require an expensive budget when scaling-out a system, NoSQL databases can vary from BASE (Basically Available, Soft state, Eventual consistency) to ACID and require less budget than SQL databases for scalability (both in hardware and software). The argument of choosing which databases for applications that run on big data, especially, social network applications has still maintained open questions. This study aims to contribute a practical discussion such questions by introducing a brief-but-concrete comparison between SQL and NoSQL databases and among four categories of NoSQL databases (Key-value, document, column, graph) under practical perspective. Moreover, the study also evaluates the operations cost on four basic operators (Insert, select, update, delete) running on a small-scale social network application with SQL and NoSQL databases. The experimental results confirm that NoSQL can be a good candidate for big data applications, especially, social network applications where select and insert operators are the utmost requirements.
Minh-Son Dao, Yen-Nhi Tran-Thi,, Van-Cuong Mai and Tuan-Anh Nguyen-Gia, 2018. NoSQL vs. SQL: Operations Cost Evaluation; A Case Study on a Social Network Application. Journal of Engineering and Applied Sciences, 13: 3278-3283.