\section{CONCLUSIONS}
\label{sec:Conclusion}

In this article, we proposed OBAS, an OLAP Benchmark for Analysis Services for the performance evaluation of analysis services based on the MDX query language and performance metrics of services. The differentials of OBAS are: (1) it assists in the performance assessment of analysis services by computing the processing costs of OLAP queries written in MDX; (2) it incorporates three performance metrics: response time, rate of execution and service reliability; and (3) it uses the metaphor of data cubes with increasing volumes of data. Using OBAS, we compared two analysis services that adopt XMLA as communication interface: the SQL Server Analysis Services (SSAS) and the Pentaho Mondrian. The results showed that OBAS can be used to compare these services with respect to performance and reliability regarding the number of threads. The extension of OBAS to enable the performance analysis of Spatial OLAP services is a future work. Also, since not all analysis services support XMLA, the incorporation of other communication interfaces (e.g. OLAP4J) can be investigated further. The performance evaluation of analysis services in the context of parallel databases or of databases stored in clouds to take the advantage of current multi-core processors or of a superior scalability, respectively, is another approach to be exploited. Finally, we also intend to evaluate queries and massive updates that occur simultaneously.