Anmerkungen:
Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
Beschreibung:
Worldwide the volume of complex unstructured and structured data is growing exponentially without any form of control. This large amount of data, defined as Big Data, improves the need for specialized Database Management Systems (DBMSs): the NoSQL DBMSs. Scalability, flexibility of modelling and ability to process multiple information are only a few advantages of NoSQL solutions. However, the variety of data results in a heterogeneous implementation of different NoSQL DBMSs with individual parameters. One of the most difficult challenges in working with NoSQL is to find out the right configuration. Wrong settings are the key to low performance, high cost and poor quality. To address these issues, various performance models, able to predict the effects of changing the runtime parameters, were built. This paper provides an overview of external and internal performance criteria, presenting how these are handled in performance models like Queueing Network and Petri Nets, analysing their scope and comparing their applicability in performance prediction. To evaluate each model, the research explains the method of benchmarking using the YCSB tool. ; publishedVersion