Cubes and Multidimensional Analysis

Microsoft SQL Server 2008 Analysis Services Unleashed Book Review Chapter 6

As I mentioned in the earlier discussion of multidimensional space, the terms used in Analysis Services are common to the industry and do not map to spacial dimensions as a physicist might expect. Specifically, the goal of Analysis Services is to limit the spacial features to a predefined subset of all possible combinations, even for the data loaded into a cube. A geometric cube mimics the universe in having theoretically infinite space, but the data in an Analysis Services cube has predefined boundaries, starting and ending points, specific attributes within dimensions, and measures. Designing a cube essentially reflects business intelligence mapped onto infinite space. I do not believe this mapping happens automatically but that the intelligent structure needs to be designed based on current and past information, and best estimates of the future.

As with chapter five, this chapter also documents sample DDL (data definition language) for a cube. These examples extend what is available in the free SQL Server Books Online. Important to this chapter is not just the cube, but also the perspective, a subset of the cube. The statement on page 72 is important for perspectives, that they do NOT act a security device to prevent access to other parts of a cube. Instead, the perspectives have a simpler task, namely to reduce the visual combination clutter presented to users through applications.

SQL Server Books Online has a good overall discussion of Planning and Architecture. This specific chapter focuses on the heart of the technology, the linkage between cubes and dimensions. As mentioned earlier, these attributes will help form the tuples to help determine a unique coordinate within the cube. Again, this book covers the architecture of Analysis Services, so if you were expecting a book which has to-do exercises, you might be disappointed. I, however, value the builder documentation which complements other literature on how to accomplish projects. This book not only describes the Analysis Services technology but also includes key terms and performance tips on making the technology work well in production.

Gorbach, I., Berger, A., & Melomed, E. (2009). Microsoft SQL Server 2008 Analysis Services Unleashed. Indianapolis, IN: Pearson Education Inc.
ISBN: 0-672-33001-6

Dimensions in the Conceptual Model

Microsoft SQL Server 2008 Analysis Services Unleashed Book Review Chapter 5

Chapter five starts with a table which maps domains (a limited list of values) from relational technology to multidimensional analysis technology. I have not used the word domain in my experience, probably because I come from a statistical and developer background, and not the DBA theory background. The word domain should mean something to a DBA.

More specifically, the discussion on page 44 claims that “implementations of relational databases essentially ignore this definition”. I am not sure all of what this claim means, but my mind quickly goes to a challenge in all relational transactional systems, namely how to define what a domain is especially when the core database methodology is built on keeping all transactions. Microsoft has acquired new technology for managing domains, called Master Data Services (I will not be commenting on what this technology does since it is beyond scope of reviewing Analysis Services). I will say now that I can imagine some complicated future design linkages between Master Data Services and Analysis Services when SQL Server is the data source.

Continue reading “Dimensions in the Conceptual Model” »