1 Introduction

Data is an increasing value resource in many contexts, for example, in a company to schedule and monitor effectively the company’s activities. Hence, a data manipulation system must collect and classify the data, by means of integrated and suitable procedures, in order to produce in time and at the right levels the synthesis to be used to support the decisional process, as well as to administrate and globally control the company’s activity. While databases are the place where data are collected, data warehouses are systems that classify the data. According to William Inmon, widely considered the father of the modern data warehouse, a Data Warehouse is a “ Subject-Oriented, Integrated, Time-Variant, Non-volatile collection of data in support of decision makin”. Data Warehouses tend to have these distinguishing features: (1) Use a subject oriented dimensional data model; (2) Contain publishable data from potentially multiple sources and; (3) Contain integrated reporting tools.

OLAP (On-Line Analytical Processing) is a key component of data warehousing, and OLAP Services provides essential functionality for a wide array of applications ranging from reporting to advanced decision support. According to [www.olapcouncil.org] OLAP “... is a category of software technology that enables analysts, managers and executives to gain insight into data through fast, consistent, interactive access to a wide variety of possible views of information that has been transformed from raw data to reflect the real dimensionality of the enterprise as understood by the user. OLAP functionality is characterized by dynamic multidimensional analysis of consolidated enterprise data supporting end user analytical and navigational activities including calculations and modeling applied across dimensions, through hierarchies and/or across members, trend analysis over sequential time periods, slicing subsets for on-screen viewing, drilldown to deeper levels of consolidation, rotation to new dimensional comparisons in the viewing area etc.”. The focus of OLAP tools is to provide multidimensional analysis to the underlying information. To achieve this goal, these tools employ multidimensional models for the storage and presentation of data. Data are organized in cubes (or hypercubes), which are defined over a multidimensional space, consisting of several dimensions. Each dimension comprises of a set of aggregation levels. Typical OLAP operations include the aggregation or deaggregation of information (roll-up and drill-down) along a dimension, the selection of specific parts of a cube (dicing) and the reorientation of the multidimensional view of the data on the screen (pivoting).

OLAP functionality is characterized by dynamic multi-dimensional analysis of consolidated enterprise data supporting end user analytical and navigational activities including:

Two main approaches to support OLAP are (1) ROLAP architecture (Relational On-Line Analytical Processing), and (2) MOLAP architecture (Multidimensional On-Line Analytical Processing). The advantage of the MOLAP architecture is, that it provides a direct multidimensional view of the data and is normally easy to use, whereas the ROLAP architecture is just a multidimensional interface to relational data and requires normally an advanced knowledge on the SQL queries. On the other hand, the ROLAP architecture has two advantages: (a) it can be easily integrated into other existing relational information systems, and (b) relational data can be stored more efficiently than multidimensional data.

Hence ROLAP is based directly on the architecture of relational databases and different venders just extend the SQL standard language in various ways in order to implement the OLAP functionality. For example in ORACLE the cube and rollup operator expands a relational table, by computing the aggregations over all the possible subspaces created from the combinations of the attributes of such a relation. Practically, the introduced CUBE operator calculates all the marginal aggregations of the detailed data set. The value ’ALL’ is used for any attribute which does not participate in the aggregation, meaning that the result is expressed with respect to all the values of this attribute.

The MOLAP architecture is basically cube-oriented. This does not mean that they are far from the relational paradigm in fact all of them have mappings to it but rather that their main entities are cubes and dimensions. In practice, the cube’s data are extracted from databases using standard SQL queries and stored in “cubes”, which beside the data extracted contain pre-calculated data in order to speed up the viewing. One of the big advantage of the cube-oriented approach is its intuitive use of generating a multitude of various “views” of the data; another advantage is speed: Having the data in a cube, it easier to reorganize the data than in ROLAP.

In this paper, the cube-oriented architecture is analyzed. In section 2 a definition of the fundamental concepts, such as datacube, is given. In section 3, the operators on datacubes are defined and analyzed. An interpretation of these operations in the light of OLAP is given in section 4. Section 5 introduces the basic concept of pivot-table, which is identified as a 2-dimensional representation of a datacube. Section 6 briefly summarizes the connection to modeling and to LPL. Finally, some reflections about spreadsheets as a modeling tool are given. See also [5], [3], [8].