WARF TECHNOLOGIES

Given the ever increasing amount of “big data” available, new analysis and processing tools are needed to fully leverage it. Big data is challenging to manage, requiring sophisticated mechanisms and parallel processing.

One problem is that many existing systems use algorithms designed for older hardware models. As hardware continues to upgrade, new software methods are needed to keep pace.

UW–Madison researchers have developed “Widetable,” a query processing method that provides a faster and more efficient means for scanning data across multiple tables.

The method accelerates queries by denormalizing multiple tables of a relational database into a smaller number of “wide row” tables using an outer join function. Such denormalization substantially increases the amount of data that must be stored to represent the database. While this might be expected to slow scan rates, speed is gained by eliminating other time-consuming operations.

• Software package
• Database management

• Tenfold query speedup
• Better scalability
• Can run on many core machines

Promising simulations have been conducted using the industry standard benchmark.

For More Information About the Inventors

• Jignesh Patel

Related Technologies

• WARF reference number P130164US01 describes a data processing method called Bitweaving that provides a faster means for scanning data contained in a single table.

Tech Fields

• Information Technology : Computing methods, software & machine learning