IBM (s ibm) is again expanding its PureSystems line of converged hardware, this built specifically for analytics. When the company debuted the first PureSystems, which combine compute, networking and storage, in April, Big Blue said it had spent three years and $2 billion developing the line.
The company’s new PureData System for Analytics features Netezza technology, which enables analytics inside databases. The New York Stock Exchange is already using the system to spot trading peculiarities that might warrant investigations, said Phil Francisco, vice president of big data product management at IBM.
“They keep track of every element of trading activity on their trading floors daily,” Francisco told me. “… And they’re able to do analysis on seven years of data.” The system also monitors systems-level data and shows if there’s enough capacity to handle major changes in trading volumes.
To capitalize on growth markets and do business with companies with lower IT infrastructure budgets, IBM is also releasing a miniature version of the PureApplication system for quickly deploying applications.
Also new is a PureFlex converged system — with compute, networking and storage all in one — targeting managed service providers. The system precludes setup and system administration, which can cut costs. It also simplifies the process of adding capacity on infrastucture in the data center, Francisco said.
IBM arrived to the converged-hardware party well after Oracle (s orcl) unveiled its Exadata Database Machine, a database appliance and several specialized boxes, including the Big Data Appliance. (Updated: IBM disputes this, claiming it was working on data warehouse appliances back in 2003.) EMC(s emc) is also in this market with the Greenplum Data Computing Appliance.
In October, following the introduction of the PureData line, my colleague Stacey Higginbotham questioned whether big data needs a specialized box. The real development, she wrote, was not the technological achievement but the acknowledgment that providing easy-to-use services is important.
This story was updated Tuesday after IBM objected to a suggestion that Oracle had beaten it to market with converged hardware, citing work in 2004 on data-warehouse appliances.