Teradata to Add Columnar Processing

Posted on September 23, 2011 By Stuart J. Johnston

RssImageAltText

Teradata yesterday announced Teradata Columnar, a database technology that integrates column and row-based databases, letting users mix-and-match columnar and row-based physical storage, whichever best suits an application.

While row-based databases are optimized to access and analyze data stored as rows, that model is not appropriate for all circumstances and can create performance bottlenecks.

In order to address that shortcoming, Teradata (NYSE: TDC) said it will ship a new component beginning in December that incorporates columnar database capabilities into the latest version of its flagship Teradata Database 14.

A columnar database helps to eliminate performance bottlenecks by storing data in columns instead of rows, according to company statements.

Using the columnar technology, when executing a query, just the data in columns necessary for processing or analysis are retrieved, a shortcut which cuts the amount of time that would be required for input and output from/to a row-based database.

A row-based database approach would instead read data from all of the columns, the statement said, putting much larger time constraints on data access.

"The flexibility of Teradata Columnar enables analytical applications to be supported by the appropriate table structure with optimized response times from a single data warehouse," Scott Gnau, president of the company's Teradata Labs, said in the statement.

The new columnar capability lets users mix-and-match columnar and row-based physical storage depending on when it best suits an application.

Among the situations that may prove to be most appropriate for using columnar storage and processing are in telecommunications customer service applications, banking customer relationship management, and retail purchasing agent activities.

"In addition, it automatically picks the best compression method and dynamically adapts the compression mechanism as data evolves over time," Gnau added.

The database automatically selects from among six types of compression, including run length, dictionary, trim, delta on mean, null and UTF8 based on the column demographics, the company said.

Stuart J. Johnston is a contributing editor at InternetNews.com, the news service of Internet.com, the network for technology professionals. Follow him on Twitter @stuartj1000.


Comment and Contribute
(Maximum characters: 1200). You have
characters left.

InfoStor Article Categories:

SAN - Storage Area Network   Disk Arrays
NAS - Network Attached Storage   Storage Blogs
Storage Management   Archived Issues
Backup and Recovery   Data Storage Archives