Kinetica Db Inc., which makes a high-speed in-memory database designed for use with hardware graphics processing unit chips, is adding programs to make its product and applications easier to install and operate.

The Install Accelerator and Application Accelerator programs “help customers translate business requirements into best practices for integrating Kinetica into their environment, designing solutions and implementing new use cases on Kinetica,” the company said in a statement.

Install Accelerator package is intended to help customer to design, build and deploy their first application using Kinetica. It includes a two-week consulting engagement and a one-year software license for up to one terabyte of memory.

The Application Accelerator package is meant to help customers quickly build and deploy any application. It includes two to four weeks of consulting and a one-year license for up to three terabytes of memory.

Founded in 2009, Kinetica has raised $13 million from a venture fund headed by former Oracle President Ray Lane for its campaign to address some of the scalability problems of key value stores like HBase and Cassandra, according to Mike Perez, vice president of services. The company’s novel architecture applies GPUs to greatly accelerate parallel processing in a column-store engine.

Databases running on standard Intel hardware can take advantage of, at most, 32 processor cores, Perez said. GPUs, which were first built for gaming applications involving high-speed video rendering, bring as many as 4,000 additional cores to the process, he said. Kinetica is targeting applications like geospatial analysis, online advertising, route optimization, Internet of things and recommendation engines, as well as anything that requires stream processing.

“GPUs are built to process millions of simple concurrent tasks, like processing multiple shades of gray or colors,” Perez said. “It’s about breaking down workloads into repeatable processes.” Kinetica can process over one billion simple operations – such as calculating minimum and maximum values in a data store – in less than a second, he said. The company claims it can speed up analytical processing for multi-billion row data sets more than 100-fold compared with leading in-memory and analytical databases.

Kinetica is a proprietary architecture that works with other data stores and visualization engines (diagram above). It works with a wide variety of open-source analytics tools like the Apache Kafka messaging engine and Apache Spark processing framework. The package supports RESTful application program interfaces as well as the Python and C++ programming languages. A SQL-92-compliant query language will be available before year’s end.

Pricing is based upon memory capacity, but Kinetica wouldn’t specify details. The company employs 50 people and has 10 customers, including the U.S. Postal Service, the U.S. Army and Pacific Gas and Electric Co.

Image courtesy of Kinetica