Some simpler and more economic models

One of the most important trends in the BI industry in recent years has been the emergence of so-called self-service BI tools such as QlikView, Tableau and Power Pivot. These tools are designed to give users the ability to create small BI solutions on a small scale with little or no help from IT departments.

Somehow, the traditional tools have always been a kind of BI tool that allows the final users to create their own queries and reports, but still require an IT professional to design and build the analytical database and the underlying data store.

This means that they are usually included with other more traditional BI tools, where database design, report presentation and access to data are strictly controlled by the IT department.

However, many organizations, especially the smaller ones, do not have the resources to conduct a large BI project, and if they try, the failure rate of such projects is usually very high. Thus, the use of self-service BI tools is attractive for a certain class of users because they allow them to analyze their information.

The traditional multidimensional BI model with dimensions and variables is getting old. This technology was designed in the era of 32-bit servers with one or two processors with less than one gigabyte of RAM, when disk storage was the only option for databases. In addition to this: a) the high costs of BI solutions (licenses, servers, man / hours in creating models and cubes, training), b) the tools are unintuitive, they need many training hours, c) they are not flexible, d) difficulty of acceptance by final users (adjustment to user requirements).

Times have changed, modern hardware is radically different today, a new generation of columnar databases (Tabular Models) based on computers’ memory, has set the standard for analytical queries. For this reason, the online analytical processing models (OLAP) should adopt this new technology to be in tune with the market

Despite the success of BI multidimensional model, there was always a perception that it is difficult to learn. Some database professionals, used to relational data modeling, struggle to learn multidimensional concepts, making the learning curve difficult, and this cause the movement from the complex world of the multidimensional model towards the relatively simple, familiar and inexpensive concepts of the Tabular model.

References

Microsoft SQL Server 2012 Analysis Services ; The BISM Tabular Model

607 total views, 3 views today