Performance Measurement Mongo Vs SQL

All the insights are derived from “Performance comparison of the most popular relational and non-relational database management systems”(Author: Kamil Kolonko)

For measurements, workloads are needed, and then these workloads are executed on both of the measured databases.

WorkLoad Definition: Before defining the workloads we need to identify the parameters for the workload.

operationcount  — number of executed operations,
readproportion   — proportion of read operations (e.g. 0.5 means 50%   
                   of total operations are database value reads), updateproportion — proportion of update operations,
scanproportion   — proportion of scan (full collection reads)                 
                   operations,
insertproportion — proportion of insert operations, readmodifywriteproportion — proportion of batches of read, update 
                            and insert operations.

Based on the above parameters and combining them in different proportions, workloads(WL) are created:

╔══════════════════╦══════╦══════╦══════╦══════╦══════╦══════╗
║                  ║ WL A ║ WL B ║ WL C ║ WL D ║ WL E ║ WL F ║
╠══════════════════╬══════╬══════╬══════╬══════╬══════╬══════╣
║ operationcount   ║ 1000 ║ 1000 ║ 1000 ║ 1000 ║ 1000 ║ 1000 ║
║ readproportion   ║  0.5 ║ 0.95 ║    1 ║ 0.95 ║    0 ║  0.5 ║
║ updateproportion ║  0.5 ║ 0.05 ║    0 ║    0 ║    0 ║    0 ║
║ scanproportion   ║    0 ║    0 ║    0 ║    0 ║ 0.95 ║    0 ║
║ insertproportion ║    0 ║    0 ║    0 ║ 0.05 ║ 0.05 ║    0 ║
║ readmodifywrite  ║    0 ║    0 ║    0 ║    0 ║    0 ║  0.5 ║
╚══════════════════╩══════╩══════╩══════╩══════╩══════╩══════╝

Below are the results of the execution

download this report here