IDC Report: Cloudera Leading Hadoop Distro Choices
In a report released by analyst firm IDC this week surveying attitudes and adoption of Apache Hadoop in the enterprise, Cloudera took top honors in leading the field as the distribution of choice.
The study, commissioned by Red Hat and titled “Trends in Enterprise Hadoop Deployments,” found that of the 32% of respondents who indicated their firms have existing Hadoop deployments, Cloudera is leading the field for adoption with nearly a quarter of all enterprise Hadoop deployments being Cloudera installations.
Per the IDC Report:
”The three leading suppliers – Cloudera, MapR, and Hortonworks – dominate the enterprise Hadoop scene. When asked about the reasons for selecting the said distributions, the vast majority of respondents cited support, management and storage costs. In other words, the “DIY” model for Hadoop appeals to only a few businesses; the rest mostly go for commercial variants, much like the route they opted for with Linux.”
As noted earlier, the survey, which surveyed 202 IT professionals responsible for management, administration, and/or decision-making, said that 32.2 % had already deployed some form of Apache Hadoop. The single largest primary use case for the framework was “analysis of transactional data from sales or point-of-sale systems” at approximately 42%. Coming in at a virtual tie at about 38% each were the following use cases:
- Analysis of online customer behavior data
- Service innovation (e.g., to aid development of a new service for clients), and
- Analysis of operations-related data
“Analysis of machine or device data” was the next use case, registering at about 32%.
The IDC study, which was limited to firms that are currently using Hadoop or considering using it, found that there is still plenty of opportunity for jockeying for position for Hadoop vendors over the next couple of years. IDC says that an additional 31.2% of respondents indicated that they have plans to deploy Hadoop within 12 months, and another 36.6% say that their deployment schedule could go beyond 12 months.
Cost may factor into the picture as enterprises consider Hadoop deployments. Nearly 50% of the respondents indicated that their firms pay anywhere from $100,000 to $500,000 in data migration costs, said the study. “Some of these costs are one-time costs associated with developing a data conversion module, while some of these costs are recurring costs associated with ongoing conversions,” said the report.
With that level of spend, tangible ROI is an important factor to consider. Approximately 45% of the respondents said that they expect returns between $1 million and $10 million in quantified business benefits, with nearly 30% saying they expect returns exceeding $10 million.
The full report can be found here.