OLAP tools are expensive investment for any company. Are there other choices for decision support tools associated with data warehouses?
For years, companies have been building general purpose decision support systems with relational databases and ad hoc query/reporting tools. Through graphical front-ends using point-and-click construction techniques, these ad hoc query/reporting tools automate the task of writing the SQL (Structured Query Language) used to access relational databases.
These tools work well for basic data retrieval and frequently provide a number of data formatting and report writing options. Their GUI interfaces expand the group of users who can directly access information in the data warehouse beyond the IS department to the technically-savvy decision makers in an organization. The main advantage of the ad hoc query tools is their endorsement of an open architecture by utilizing SQL standards. Conforming to this industry standard allows ad hoc query tools to readily adapt to changing information systems requirements.
While these report writers offer many advantages as basic reporting tools, they fall short of meeting many of the requirements organizations place on their decision support systems. Their greatest shortcoming results from the fact that end-users are still directly involved in the SQL definition process. In order to achieve this simplified ad hoc query environment, reports writers have had to sacrifice analytical complexity. To perform analyses on micro-segments of the data warehouse, a query plan would need to consist of several related SQL statements. This SQL execution plan is difficult for even experienced database professionals to write, and impossible to construct with single-statement relational report writers. As a result, relational report writers do not provide the complex analyses required by today's business users.
Vendors of OLAP products are classified as either multidimensional OLAP or relational OLAP based on the underlying architecture of the system. The premise of ROLAP is that OLAP capabilities are best provided directly against the relational database, i.e., the data warehouse. ROLAP systems leverage relational databases to provide multidimensional analyses in a three-tier architecture. The relational database handles data storage requirements, and the ROLAP engine provides the analytical functionality.
Relational OLAP tools provide a solution by combining the direct data access functionality of the ad hoc query/report writers with the analytical sophistication of the multidimensional OLAP tools. Relational OLAP provides an open, scaleable architecture that meets all the requirements for decision support. The ROLAP architecture is better equipped to handle large numbers of dimensions than the MD-OLAP architecture. In addition, by providing a larger operational envelope in all cases, ROLAP is better equipped to deal with fluctuating query-demand profiles. The ROLAP architecture is better equipped to handle large data volumes than the MD-OLAP architecture.
Relational OLAP depends on the rational data models of the data warehouses to access the data. Therefore, data is not stored in the multidimensional model as it is with OLAP. The timing of queries can be significantly larger with ROLAP due to the translation that the ROLAP tools have to performed from the relational model to the multidimensional model that the user is looking for.
Data mining is the data-driven extraction of information from large databases. It is the process of automated presentation of patterns, rules or functions to a knowledgeable user for review and examination. Here the human plays an essential role in the paradigm because it is only he, the analyst, who can decide whether a pattern, rule or function is first interesting, second relevant and third useful to the enterprise.
In business and in the press, data mining is hot. As with most waves that flash through our culture, care must be taken to separate the fact from the fantasy. Data mining is a useful tool, a new approach which combines discovery with analysis. Data mining is not a newly discovered branch of mathematics embodied in software that will, when hooked-up to a large and problematical database, inexplicably and inevitably reveal the business insights contained in the millions of records stored therein. Yet still it is important. It is an area that will increasingly become mandatory for competitive businesses.
www.access.digex.net/~grimes/olap/olap.collections.html