Performance Evaluation of SQL and No SQL Database Management Systems in a Cluster

摘要

In this study, we evaluate the performance of SQL and No SQLdatabase management systems namely;Cassandra, CouchDB, MongoDB, PostgreSQL, and RethinkDB. We use a cluster of four nodes to run thedatabase systems, with external load generators.The evaluation is conducted using data from TelenorSverige, a telecommunication company that operates in Sweden. The experiments are conducted usingthreedatasets of different sizes.The write throughput and latency as well as the read throughput and latencyare evaluated for four queries; namely distance query, k-nearestneighbour query, range query, and regionquery. For write operations Cassandra has the highest throughput when multiple nodes are used, whereasPostgreSQL has the lowest latency and the highest throughput for a single node. For read operationsMongoDB has the lowest latency for all queries. However, Cassandra has the highest throughput for reads.The throughput decreasesas the dataset size increases for both write and read, for both sequential as wellas random order access. However, this decrease is more significant for random read and write. In thisstudy, we present the experience we had with these different database management systems including setupand configuration complexity.