Parker et al [1] studied the performance difference between MongoDB and Microsoft SQL Server on basis of a number of insert, update and select scenarios. As the result of their study, they conclude that MongoDB has got a better performance when it comes to insert, update and simple queries. However, Microsoft SQL Server performs better when updating and querying the non-key fields (attributes) as well as for aggregate queries. In this Master thesis we replicate the mentioned study in order to validate its results. Moreover, we extend the study with more data in order to monitor how both databases perform when the data grows. Although there are differences between the exact values in the results of both studies, we also conclude that MongoDB performs better in case of insert, update and key field related queries. However, when non-key field related queries are utilized we see that SQL Server has got a better performance. In case of aggregate queries we also see the similar results as in [1] i.e. SQL Server performs better than MongoDB. However, we see that the new aggregate framework of MongoDB provides a significant improvement with respect to the method used in [1] which results in a slightly better performance than SQL Server. Furthermore, we show how different implementation of the experiment can significantly affect the results. The results of this study will help us to answer the question whether or not to choose a NoSQL database in case of a modest-sized database with structured data from the performance point of view.