CST(47)-OLAP and Multidimensional Data

Online Analytical Processing (OLAP) tools create multidimensional data views on top of ordinary 2-D SQL databases (or using specialized multidimensional databases). OLAP’s multidimensional access lets you formulate more sophisticated queries, and then look at the results accordingly. Think of OLAP as a multidimensional spreadsheet with multiple axes. For example, you could have a product database they you access via multiple dimensions such as time, region, customer, store, price, and sales. The idea is to let you explore data using different dimensions.

OLAP’s multidimensional model makes it easier to visualize data. Instead of navigating through multiple tables and rows, you look at data through multidimensional views. Relational/OLAP tools provide this function by introducing a layer of abstraction on top of SQL databases that hides the physical structure of normalized relational tables. Instead, you get to see multidimensional views of that same data. OLAP obviously provides a more intuitive way to look at data, especially if you used a spreadsheet’s pivoting feature.

The OLAP model comfortably handles data in ten or less dimensions. Beyond this, servers fail from index overload. In addition, our brains may also fail from n-dimensional visual overload. The OLAP tool vendors are split into three camps:

1. Relational OLAP tools. These are OLAP client tools that create multidimensional views by extracting data from ordinary SQL databases. These tools simulate multidimensional visualizations using sophisticated indexing, caching, and Meta data techniques. Tools in this category include Brio’s DataPivot, Business Object’s Mercury, Cognos’ PowerPlay, Andyne’s Pablo etc.

2. OLAP-savvy SQL servers. The vendors in this camp offer specialized SQL databases that are optimized for running OLAP-savvy queries. The specialized SQL database vendors obtain performance gains by using special indexing techniques; parallel joins, OLAP-savvy SQL generators, and SQL query extensions that help the database engine optimize the query. Vendors in this category include Red Brick’s Red Brick Warehouse and MicroStrategy’s DSS/ Server.

3. Multidimensional DBMSs (MDBMSs). Rather than storing data as keyed records in tables, MDBMS provide specialized database engines to store data in arrays along related dimensions called hypercubes. Most MDBMSs use indexing-intensive schemes to optimize access to these cubes. MDBMS vendors typically provide OLAP client tools that are optimized for their engines – there are no standards for MDBMS data access. OLAP/MDBMS tools include Arbor’s Essbase, Kenan’s Accumate, Pilot’s Lightship Server, Oracle/IRI’s Express, SAS OLAP++, etc.
CST(47)-OLAP and Multidimensional Data Reviewed by 1000sourcecodes on 21:40 Rating: 5
Powered by Blogger.