Hadoop Big Data Use Emerging But Unsettled

Register now

July 28, 2011 – Under increasing mounds of data, IT and business are warming to Hadoop, though sporadically and in addition to existing infrastructure, according to new findings from Ventana Research.

Ventana’s report, “Hadoop and Information Management: Benchmarking the Challenge of Enormous Volumes of Data,” outlines enterprise attitudes toward Hadoop adoption with large-scale data sets.

About one-third of respondents (34 percent) are using or plan to use some form of Hadoop within 12 months, and 19 percent are evaluating their interest in Hadoop or are planning to adopt use within the next two years, the report stated. When used, only 37 percent of organizations replace another technology in their IT infrastructure with Hadoop, more often accessing it in addition to the existing stack, particularly for unstructured or event data, the report stated.

Overall use of Hadoop is emerging but “unsettled,” as seen in responses to differing application of the technology’s iterations, according to the report. Apache Hadoop was the dominant distribution of the technology, though 55 percent of those working with Hadoop use five or more components like MapReduce for big data needs. The report indicated that this also opens opportunities for commercial vendors to tie together functionality and services surrounding Hadoop.

Few IT users marked themselves as “very satisfied” with Hadoop for large data sets (30 percent), with 44 percent of business users finding the same level of satisfaction in the survey. Ventana remarked those levels – as well as big data’s connection to virtualization – are representative of a vastly expanding but “immature” market.

David Menninger, VP and research director at Ventana, notes that virtualization and advances in capabilities with large data sets are linked in adoption rates and interest, especially with an increase in the most prevalent form of big data processing, clustering. The report pointed to expectations of Hadoop clusters to grow beyond 10 nodes, and interconnected computers to process that clustering to be widely involved in the mass move to cloud computing activity.

“The survey results show cluster sizes growing and more organizations using offsite – hosted or SaaS/cloud – in the future,” says Menninger. “It would seem that these trends are converging and will make it easier [and] cheaper to create large clusters.”

The report is based on survey answers from more than 160 executives and IT officers from industries of varying sizes involved in finance, insurance, real estate, sales, manufacturing, and government and education, largely based in the U.S. and Asia. Information Management was a media co-sponsor of the report.

To access the report for a fee, click here. For a Web seminar at 1 p.m. ET Thursday based on the findings, register here.

For reprint and licensing requests for this article, click here.