Authors:
Anikó Vágner
and
Mustafa Al-Zaidi
Affiliation:
Department of Information Technology, Faculty of Informatics, University of Debrecen, Hungary
Keyword(s):
NoSQL Databases, Cluster, Replication, Sharding, Distribution Models.
Abstract:
One of the reasons that NoSQL databases were born is that they can be used in clusters, namely, computers work together, share data and from the client side, the clustered computers look as if there is only one computer. In this paper, the distribution models of two NoSQL databases are introduced. We chose the databases from the database ranking website (noa, 2023a), exactly the two first NoSQL databases: MongoDB and Redis. However they belong to two different NoSQL database categories, they use similar key-value pairs which is the main basis of the clustering. Additionally, the distribution models do not depend on the categories of the databases, both database management systems know sharding and master-slave replication, and can use these two distribution models together. These two database management systems do not know peer-to-peer replication. Our goal was to get to know whether there are similarities between the structures of clustered computers of each database management syst
em. If we consider the theory, the answer should be yes, they are similar to each other: for the sharding 2 computers are enough, similarly for the replication 2 computers are also enough, and if both of the techniques are used, 4 computers should be enough.
(More)