Optimization and Tuning in Data Warehouses

This is an excerpt from the content


Optimization and tuning in data warehouses are the processes of selecting adequate optimization techniques in order to make queries and updates run faster and to maintain their performance by maximizing the use of data warehouse system resources. A data warehouse is usually accessed by complex queries for key business operations. They must be completed in seconds not days. To continuously improve query performance, two main phases are required: physical design and tuning. In the first phase, data warehouse administrator selects optimization techniques such as materialized views, advanced index schemes, denormalization, vertical partitioning, horizontal partitioning and parallel processing. Generally, this selection is based on most frequently asked queries and typical updates. Physical design generates a configuration Δ containing a number of optimization techniques. This configuration should evolve, since data warehouse dynamically changes during its lifetime. Thes ...