The purpose of this article is to provide a comparisson of relational and non-relational databases in terms of performance, scalability, availability, expressiveness, consistency, manageability, integrability and flexibility.
There are several studies done to test the performance of relational and non-relational databases. The tests are performed doing CRUD operations on the different databases and verify the results using various volumes of data. Also, performance depends on the implementation of the database. Based on Gyorodi et al. (2015), MongoDB (No-SQL DB) performed better on INSERT, UPDATE and DELETE operations than MS SQL Server but the performance was lower on SELECT operations. Here are some tables that show the comparison.
Relational databases scale up vertically, what it means that it needs to add more power to the hardware such as increasing RAM, adding more CPU and disk. On the other hand, non-relational databases scale up horizontally, so it is possible to increase the capacity by adding more servers (Faraj et al., 2014).
Non-relational databases offer more availability than relational databases. Servers can be added or removed without any downtime. Most implementations of non-relational databases support data replication, storing data in different clusters or datacenters to ensure high availability (Serra, 2015).
Relational databases use Structured Query Language or SQL to interact with the data, and it is compelling. On the other hand, non-relational databases do not have any standard to query data, some of them support UnSQL (unstructured query language), but most of them depend on the implementation of the database. For example, Cassandra uses CQL (Cassandra Query Language), or MongoDB uses Mongo Query Language (Faraj et al., 2014).
In relational databases, all users see the same version of the data after a transaction is committed in the database. Instead, in non-relational databases, consistency is eventual, so there is no guarantee for reading and writing after a transaction is performed that will be immediately consistent. Hence, relational databases provide better consistency than non-relational databases (Faraj et al., 2014).
From an administration perspective, non-relational databases are extraordinarily user-friendly and do not require a database administrator like relational databases need (Nayak et al., 2013).
In non-relational databases, RESTful interfaces such as HTTP POST, PUT, DELETE are used in combination with a different standard format like JSON. Also, they provide data manipulation APIs. Data manipulation in relational databases is done through DML or data manipulation language (Faraj et al., 2014).
Relational databases store data into rows that are predefined tables with different columns. The data has to fit into those structures, but in non-relational databases, there is no schema, so it is possible to store unstructured data with no limitations on the types of fields. This functionality allows non-relational databases to provide higher flexibility than relational databases (Faraj et al., 2014).
Gyorödi, C., Gyorödi, R., & Sotoc, R. (2015). A comparative study of relational and non-relational database models in a Web-based application. Int J Adv Comput Sci Appl, 6(11), 78-83.
Li, Y., & Manoharan, S. (2013, August). A performance comparison of SQL and NoSQL databases. In 2013 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM) (pp. 15-19). IEEE.
Faraj, Azhi & Rashid, Bilal & Shareef, Twana. (2014). COMPARATIVE STUDY OF RELATIONAL AND NON- RELATIONS DATABASE PERFORMANCES USING ORACLE AND MONGODB SYSTEMS. INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING & TECHNOLOGY. 5. 976-6367.
Serra, J (2015) Relational Databases vs Non Relational Databases. Available at: https://www.jamesserra.com/archive/2015/08/relational-databases-vs-non-relational-databases/ (Last Accessed: 17 August 2019) Nayak, A., Poriya, A., & Poojary, D. (2013). Type of NOSQL databases and its comparison with relational databases. International Journal of Applied Information Systems, 5(4), 16-19.