Teradata Introduces In-Memory Technology

Teradata Corporation introduced Teradata Intelligent Memory, a database technology.

Intelligent Memory creates ant extended memory space beyond cache that increases query performance and enables organizations to leverage in-memory technologies with big, diverse data. These innovations help results-oriented organizations to use all their data to know more about their customers and business – so they can do more to drive profitability.

"The introduction of Teradata Intelligent Memory allows our customers to exploit the performance of memory within Teradata Platforms in a new way, and extends our leadership position as the best performing data warehouse technology at the most competitive price," said Scott Gnau, president, Teradata Labs. "Teradata Intelligent Memory technology is built into the data warehouse and customers don’t have to buy a separate appliance. Additionally, Teradata enables its customers to configure the exact amount of in-memory capability needed for critical workloads. It can be difficult, expensive, and impractical to keep all data in memory, and Teradata’s unique approach means the right amount of memory can be applied to the right set of data for blinding performance – automatically."

Intelligent Memory is another solution that fits into the overall Unified Data Architecture strategy, which leverages Teradata, Teradata Aster, and open source Apache Hadoop. Data in Apache Hadoop that is frequently used can be accessed through Teradata SQL-H, and based on temperature of the data, moved to Intelligent Memory to take advantage of its HPC capability.

Some organizations have opted to store all data in memory to speed answers to queries, but the cost is prohibitive and there are limits on memory capacity, which is important with the volumes of big, diverse data needed for analysis. Intelligent Memory manages the data in this new space by predictively placing the very hottest or most frequently used data into memory, then automatically updating and synchronizing it. The process doesn’t need any human intervention. It never requires that changes be made to applications, SQL queries, or how the data is stored. The company offers customers the benefits of data analytics on data in-memory. Access to data in-memory eliminates disk I/O bottlenecks and query delays, and increases system throughput Access to data in-memory is three orders of magnitude faster than access to HDD drives.

Intelligent Memory is available for any of the current Teradata workload-specific platforms running the Teradata Database. For organizations looking for departmental solutions, Integrated Data Warehouse or Active Data Warehouse, Intelligent Memory delivers the performance required to deliver analytics to drive competitive advantage.

Intelligent Memory uses algorithms that automatically age, track, and rank data to ensure effective data management and support for user queries. Inside, data can be stored and compressed in columns and rows, which maximizes the amount of data in the memory space. It places only the hottest data to the new extended memory space. Organizations benefit by being able to access the most current data quickly from system memory to satisfy the vast majority of their queries, which also achieves a better ROI.

Organizations that are faced with heightened competition and economic conditions continue to demand more from their data warehouses to help make critical business decisions. There is a need to have as much data available as possible to broaden and deepen the analyses conducted by the business. However, data warehouse performance is impacted by the rapid growth of data volume, complexity, variety, variability, and the analytic applications that use the data. Database administrators cannot manually tune and adjust settings to accommodate every change in workload.

"Teradata’s new in-memory architecture is integrated with its management of data temperature," said Richard Winter, CEO, Winter. "This is very significant, because the hottest data will migrate automatically to the in-memory layer – Teradata Intelligent Memory; the next hottest data will move automatically to solid state disk; and, so on. Teradata also provides the column storage and data compression that amplify the value of data in memory. The customer sees increased performance without having to make decisions about which data is placed in memory."

Teradata takes a systems-level approach to multi-temperature data management to ensure optimum performance to meet the service level demands of business users. To meet users’ demands, Intelligent Memory extends multi-temperature data management into the memory layer. It supplements other multi-temperature features including Teradata Virtual Storage which continuously migrates data to the appropriate level of storage, SSD and HDD technology, based on its to achieve optimum data warehouse performance.

Teradata Virtual Storage provides an automated lifecycle management process where ‘cold’ or the least used data is migrated to less expensive drives. In addition, when data is identified as ‘cold,’ or the least used, the Teradata compress-on-cold feature automatically compresses the data to deliver space savings. This process eliminates the frequent compression and decompression of hot data, is effortless for customers, and provides an increase in the amount of data stored. When user data is compressed and it is transferred to the new memory space, Intelligent Memory is able to maximize the amount of data in the memory space.

Intelligent Memory will be available as part of Teradata Database 14.10 for release in the second quarter of 2013.

Video on In-Memory
To read the article, Big Data and In-Memory: Are They Related?, by Robin Bloor