Follow Datanami:
February 28, 2014

MapR’s Hadoop Offering the Strongest, Forrester Says

Alex Woodie

Among Hadoop distributors, MapR Technologies’ product offering is the strongest, Amazon Web Services has the biggest Hadoop market presence, and IBM and Pivotal Software have the best market strategy, according to a new Forrester Wave report that slices and dices the Hadoop market and analyzes the various Hadoop product offerings and business strategies.

MapR Technologies’ Hadoop product line is tops in the industry and carries a 4.25 rating out of 5, according Forrester Research. Following closely behind are the usual Hadoop suspects, including Hortonworks with a 4.14 product rating and Cloudera at 4.04. Microsoft had the lowest score, at 2.94, while Amazon, Intel, IBM, Pivotal, and Teradata occupied the middle with scores between 3 and 4.

Forrester cited MapR‘s strong Hadoop architecture and credited it with making “unique innovations” to its M3, M5, and M7 products, including support for the Network File System (NFS), the capability to run arbitrary code (such as C++ applications, not just Java), its HBase performance tweaks, and high availability features.

MapR took the good news in stride. “We believe the fact that MapR achieved the highest score for its current offering across all vendors’ solutions is a great acknowledgement of how MapR ensures production success for Hadoop,” MapR’s chief marketing officer Jack Norris said in a statement.

The main concern with MapR is its market awareness. “Forrester clients often ask about Cloudera and Hortonworks–but not about MapR Technologies,” Forrester analysts Mike Gualtieri and Noel Yuhanna wrote in yesterday’s report, which MapR is making available on its website. “MapR Technologies has a leading solution; it must now make more noise in the market and accelerate its partnerships and distribution channels.”

Courtesy: Forrester Research

IBM and EMC Hadoop spinoff Pivotal were top Hadoop vendors in terms of strategy. The two vendors pulled in top scores in product roadmap, customer support, and ability to execute.

Forrester said that IBM (with more than 100 BigInsights Hadoop customers) benefits from analytic tooling, a global presence, and implementation services, which combine to create “a complete big data solution.”

Pivotal, meanwhile, is differentiated by its Hadoop appliance, its capability to leverage EMC engineers, and HAWQ, its SQL engine. With less than 100 Hadoop deployments, Pivotal’s customers are primarily smaller and midsize organizations.

AWS and its Elastic MapReduce (EMR) offering carried the top market presence among all Hadoop vendors. “Its EMR has already achieved considerable adoption and had the highest score for market presence in this evaluation,” Gualtieri and Yuhanna wrote. The analysts cited a very strong roadmap for AWS, including integration with Kinesis for stream processing, better integration with Redshift, support for additional NoSQL databases on Hadoop, new autoscaling features that will resize clusters based on policies, and integration with third-party business intelligence vendors.

Forrester cited Cloudera‘s willingness to listen to customers and build what they need, such as Imapla for SQL processing and Enterprise Manager for Hadoop monitoring. “Cloudera’s approach to innovation is to be loyal to core Hadoop but to innovate quickly and aggressively to meet customer demands and differentiate its solution from those of other vendors,” the Forrester analysts wrote. With more than 200 customers–some of which are storing multiple petabyte on 1,000-plus node clusters–Cloudera has one of the biggest Hadoop customer bases (second to Amazon, whose customer based is unknown).

Forrester cited Hortonworks‘ commitment to open source and its strong partner network. “Hortonworks’ strategy is to drive all innovation through the open source community and create an ecosystem of partners that accelerate Hadoop adoption among enterprises,” the analysts write. If you like the Ambari cluster management product, you can thank Hortonworks, which employs most of the committers. Hortonworks has the strongest partner network, which includes SAP, Teradata, Microsoft, and Red Hat.

Teradata is best known for its enterprise data warehouse, but it’s moving into Hadoop, too. On a technical level, Forrester cited its Hadoop management tools and SQL-H, which lets customers query data residing in its traditional Teradata warehouse and Hadoop. While it has fewer than 100 customers, “Teradata’s extensive financial, technical, and management resources can create a unique, high-performance Hadoop appliance that few other vendors can match,” Forrester says.

Forrester also rated the Hadoop products of two other IT giants, Intel and Microsoft. Intel was late to the Hadoop game, but the analysts expect Intel’s roadmap to improve over the next year. Intel’s strengths lay in its capability to leverage its ownership of Xeon chips to boost performance, as well as its security enhancements, support for Lustre file systems, and graph analytics.

Microsoft’s Hadoop strategy is closely tied to Hortonworks, with whom it has a partnership to develop a version of Hadoop that runs on Windows. One of Microsoft’s big strengths is how it runs its HDInsight Hadoop distro on the Azure cloud. Forrester also noted how Microsoft contributes to various Hadoop projects, including Hive, and how it enables SQL Server customers to execute Hadoop queries with Polybase. Going forward, Microsoft has certain things going for it with regards to Hadoop, particularly with the interplay of various tools, including SQL Server, Azure, Excel (especially PowerPivot), collaboration, and application development tools.

2014 is going to be a very active year in Hadoop, with new implementations and expansions of prior projects. “The Hadoop buying cycle is on the upswing, and the Hadoop vendors know it,” Forrester analysts write. “Pure-play upstarts must capture market share quickly to make venture investors happy; stalwart enterprise software vendors must avoid being disintermediated; and cloud vendors must make their solutions cheaper.”

Related Items:

Big Data Market Growth Exceeds Expectations

Rethinking Real-Time Hadoop

When to Hadoop, and When Not To