Teradata Merges Big Data Into Warehouse Appliance

Data analytics infrastructure provider Teradata, which released its first Aster-based database and a new MapReduce “big data” implementation two weeks ago, announced that the fifth generation of its data warehousing appliance using both components will be available in a few months.

The company made the announcement at its annual Partners Conference here at the Convention Center, a five-day international event that attracted about 3,500 attendees and continues through 6 October. All three products can be deployed separately or together as needed.

Defining Big Data For Warehouses

Teradata acquired Aster Data in March 2011 and is ramping up development of its products, including the database and its MapReduce programming model. As a result, Teradata is moving noticeably faster up the ladder in competition with companies such as Oracle, IBM, Microsoft, EMC and HP.

Teradata Data Warehouse Appliance 2690, naturally, is bigger, faster and better – as most new versions of IT hardware and software tend to be. This is a fully-integrated system that doubles the performance and triples the data storage-processing capacity of its predecessor, the company said.

Teradata defines big data as being a voluminous mix of structured and unstructured data involving complex inter-relationships that do not lend themselves to analysis with today’s traditional techniques. This makes capturing, storing, managing and analysing it extremely difficult.

All the new Teradata products are aimed at simplifying those tasks by using an intuitive interface that can be navigated by a savvy business person – not necessarily a database administrator.

“We think that by using this system is the simplest way to bring big data to the business,” Stephanie McReynolds, director of Aster Data product marketing, told eWEEK. “There’s a lot of chatter about challenges right now around big data in the enterprise involving bringing MapReduce into the enterprise, because of the complexities and finding the skills sets to support MapReduce development.”

MapReduce is a software framework introduced by Google in 2004 to support distributed computing on large data sets on clusters of computers. Parts of it are open source, parts are patented by Google. A number of companies are taking the open source portions of MapReduce and building their own analytics implementations.

“This (database, MapReduce implementation and appliance) is a solution that simplifies the business adoption of MapReduce,” McReynolds said. “The Aster Database has a framework that marries SQL and MapReduce together – this is an extension of the SQL/MapReduce framework that we have had in Aster Database for some time. It allows business people to speak (the language of business analytics is really SQL) to execute MapReduce code without having to get coders involved.”

The appliance is an integrated analytic platform that can be configured from as few as 2TB up to 315TB of uncompressed user data per cabinet, the company said. Performance-wise, it can scan data at more than 38GB per second, per cabinet.

New Incremental Levels of Performance

When the new Data Warehouse Appliance comes out early next year, it will take advantage of Teradata Columnar, which will deliver incremental levels of performance and compression, Randy Lea, vice president of products and services marketing for Teradata, told eWEEK.

The Data Warehouse Appliance reaches up to 60 percent less energy and 50 percent less floor space for the same capability as the previous generation, Lea said.

“Teradata achieved these savings by internal streamlining of all system management functions through virtualisation technology and a compression engine that provides more performance and data storage on a cabinet-to-cabinet level when compared to the previous generation,” Lea said.

Green benefits in a data centre have become a critical factor, since it is widely known that the costs for data centre power, cooling, and floor space will soon outpace the cost of the equipment in the data centre.