Hadoop Market is Neck and Neck, Forrester Says
If you’re shopping for a Hadoop distribution on which to hang your big data hat, you have your work cut out for you, according to Forrester Research, which found four strong performers in a market it says is neck and neck.
“Choosing a Hadoop distribution will be difficult for most AD&D [application development and delivery] pros who carefully consider each of these Leaders,” write Forrester analysts Mike Gualtieri and Noel Yuhanna. “Forrester doesn’t think there is a wrong choice among the Leaders in this evaluation. This is still a neck-and-neck market.”
In its Forrester Wave for Big Data Hadoop Distributions, which was released yesterday, the company put Cloudera, Hortonworks (NASDAQ: HDP), IBM (NYSE: IBM), and MapR Technologies in the leader’s sector, with little differentiation among them. It put Pivotal Software, which last year aligned itself with Hortonworks and the controversial Open Data Platform Initiative (ODPi), in the strong performer’s sector.
The company gave Cloudera, MapR, and IBM high marks for their current Hadoop offerings. Cloudera, in particular, outscored rivaled Hortonworks in the security, data, and data governance categories. MapR had the highest score when it comes to architecture, but scored the lowest in the development category. Both Cloudera and Hortonworks scored well when it came to administration.
Hortonworks narrowly edged out each of its Hadoop rivals in the strategy department. Forrester considers the acquisition and pricing strategy of Hortonworks and MapR, in particular, to be superior to the other three Hadoop players. Pivotal had the lowest strategy score, while IBM was the only vendor with a perfect score in the implementation support department.
In terms of market presence, the race between Cloudera and Hortonworks was too close to call, while the three other players lagged behind the two perceived front-runners.
Forrester lauded Cloudera for being a trend-setter in the space. In particular, it cited Cloudera’s moves in areas like SQL-on-Hadoop, and the addition of enterprise-level capabilities (like security, high availability, and governance) into the solution. Customers like Cloudera’s add-on tools, Cloudera Manager and Cloudera Navigator, Forrester says.
Hortonworks’ strength lies in its adherence to the open source Apache Hadoop product and how it seeks to form an “inclusive and broad” open source community. Forrester notes that Hortonworks is not shy about doing acquisitions to fill gaps in the product lineup, and states that customers like Hortonworks’ open source approach to innovation.
MapR is differentiated by how it delivers “extreme performance and reliability at scale,” Forrester says. In particular, it mentions MapR’s file system, which augments the HDFS API with additional read/write capabilities. Most MapR customers are looking to build big Hadoop clusters and want to use MapR’s NoSQL and streaming analytic capabilities, Forrester says.
IBM’s strength lies in its ability to deliver end-to-end advanced analytics, Forrester says. Big Blue got kudos for its work with Apache Spark, as well as a host of add-ons that plug into its BigInsights distribution, such as BigSQL, BigQuality, BigIntegrate, and InfoSphere Big Match.
While Pivotal’s Hadoop offering and strategy trails the other four, it still may be the right choice for some customers, according to Forrester, which noted that many Pivotal Hadoop customers are already Greenplum customers. Pivotal does have a strong open source strategy, Forrester noted, particularly around Greenplum, Apache Geode, Apache HAWQ, and Apache MADlib.
The analysts, Gualtieri and Yuhanna, are bullish when it comes to Hadoop adoption, and state that 100 percent of large companies will eventually adopt a Hadoop solution, which they define as including related technologies, including Apache Spark.
Forrester intends to deliver a separate report on the cloud-based Hadoop market soon. That report would feature the offerings of companies like Amazon AWS (NASDAQ: AMZN), Microsoft (NASDAQ: MSFT), and Altiscale, among others.