ParAccel Introduces High-Speed Columnar Database for Analytic Processing

Database offers dramatic speed and efficiency, can scale for critical business applications

ParAccel, Inc. has launched its first product, ParAccel Analytic Database, a high-speed, columnar database management system for decision processing such as traditional data warehousing and analytics. The new company says its product combines innovations with a “columnar, compressed, massively parallel, all-in-memory-capable architecture to yield significant gains over other database management systems in both performance and price/performance.”

The database is available as enterprise software or as a virtual or packaged appliance and runs on standard hardware from major vendors; it can be implemented as a stand-alone database or as a drop-in accelerator to improve analytic processing for Microsoft SQL Server or Oracle platforms.

ParAccel Analytic Database supports several open standards (for example, ODBC and SQL 92) and works with BI tools from Business Objects, Cognos, Microsoft and MicroStrategy.

“The data warehousing market continues to evolve towards requirements for higher-performance at affordable price points,” said Dan Vesset, IDC vice president, Business Analytics, in a statement. “For several years the demand for such DW system characteristics has outpaced the supply. The market is simply waiting for more options in the form of innovative solutions, such as the ParAccel Analytic Database, which are targeted directly at the business analytics market.”

The company notes that business intelligence and analytic workloads “typically include complex database operations such as joins, scans, aggregations, and sorts which are difficult for row-wise databases to perform without manual tuning techniques such as data partitioning, materialized views, indexes, etc.” Such complexity can be costly and can delay applications implementations.

ParAccel’s uses massively parallel compute “nodes” (rack-mounted servers or blades) behind one or more leader nodes running on standard hardware. The software’s appliance-ready design eliminates the performance tuning and maintenance typically needed for business intelligence apps.

“The ParAccel Analytic Database doesn’t require indexes or other redundant structures, so what is often characterized as a 20TB data warehouse may only contain 5TB of user data,” said ParAccel CTO Zane. “The product then compresses this data to approximately 1TB, making it feasible to run completely in-memory for extremely high performance.”

The ParAccel Analytic Database software is available now; per-gigabyte and per-note licensing, plus subscription licensing, are available. More information can be found at http://www.paraccel.com.

About the Author

James E. Powell is the former editorial director of Enterprise Strategies (esj.com).