As SSBI tools evolved, data scientists were still wrestling with the overall challenge of ﬁnding an analytical database as ﬂexible for analytics as relational databases were for transactional data processing.
Development of Analytical Databases
Progressive software vendors sought to overcome the limitations of data warehouses, cubes, and SSBIs and so they began working towards creating databases that were both ﬂexible and able to process analytical workloads. These analytical databases, or column stores, were the next step in the trend to provide business analysts the tools and ﬂexibility they need. These analytical databases have evolved into Massively Parallel Processing (MPP) analytical databases that are more ﬂexible and more performant than Cubes, even in cases where large amounts of data are being stored and queried.
However, these analytical databases require that data be copied into them using processes that are very similar to the aforementioned ETL processes, and they have similar drawbacks. The load processes are typically slower than in traditional data warehouse-based or row-based technology because there is an extra step required to optimize the data for quick analytical retrieval. This extra step is required to convert the data from a row-based format into a columnar format and then apply ﬁeld-level data compression. Although these extra steps provide signiﬁcant performance improvements, they also require additional time that delays the analysts’ ability to analyze the data. It is impossible to access real-time data in analytical databases due to this load-time latency.
Advantages and Disadvantages of Analytical Databases
Would you like to know more about SSBI and different data integration tools? Check out our free ebook, “Beyond the Data Lake”. It gives you a brilliant overview and enhances your knowledge of this specific topic.