Customers are confused about the business models, and particularly the Hortonworks model, in the Hadoop market, writes Wikibon Principal Research Contributor Jeff Kelly in “Hadoop Pure-Play Business Models Explained”. The main source of this confusion comes from a misunderstanding of where Hortonworks makes its money. As a pure-play vendor supplying pure Apache Hadoop with no proprietary additions, Hortonworks does not charge licensing fees. Instead, it makes its money providing two kinds of services:

• Subscription services including Phone and Web support and access to patches, updates and upgrades as they are released by the Apache community, and,

• Educational and professional services.

Today it gets 70 percent of its revenues from the former and the remaining 30 percent from the latter. The confusion lies in the common misconception that Hortonworks is mainly a consulting company. While it does do some consulting, and while that is definitely a profitable business, it is difficult to scale. Actually it is mainly a subscription service company, and its goal is to get 80 percent of its revenues from its subscription services.

To grow those services, Kelly writes, it is aggressively expanding its channel, mainly through reseller agreements with major enterprise software vendors such as Teradata, Microsoft and SAP. Its challenge in the coming year is to incentivize its partners’ sales forces to sell its low-margin support services rather than more lucrative, and in some cases competing, proprietary software licenses. However, Hortonworks certainly can built a successful business around annual subscriptions; the key is execution and patience as its low-margin support business puts price pressure on competitors.

Cloudera

The pioneer in this market, Cloudera provides an enterprise data hub, the “Cloudera Distribution Including Apache Hadoop (CDH)”. This includes a 100% Apache Hadoop distribution with proprietary add-on modules designed to extend the functionality of its platform. These modules include “Impala” for SQL-like interactive queries, “Search” for Google-like search functionality and “Navigator” for data lineage, access management and auditing capabilities. Its goal is to build a comprehensive data management and analytics platform on top of Hadoop, rather than waiting for the open source community,in which its engineers play an active part, to build that platform. This, writes Kelly, “is both daring and risky, with a high risk/reward quotient.”

This approach puts it in competition with much larger vendors including IBM and Pivotal. It also opens itself to customer concerns about vendor lock-in, a sore subject for many enterprises that are “paying through the nose” to Oracle and other proprietary database vendors.

Until recently Cloudera was very vocal about the open source Apache nature of CDH. That seems to have changed in October with CSO Mark Olson’s publication of an article on LinkedIn detailing the central role Cloudera’s proprietary components now play in the vendor’s strategy.

Cloudera has built a robust platform and won a strong group of enterprise customers. It has a more diverse and higher margin set of revenue streams than Hortonworks. If it can sell and execute on its Enterprise Data Hub vision it can pull away from the pure-play Hadoop competitors, but if it does it will find itself facing IBM and Pivotal.

MapR

MapR is building its business by focusing on enterprises moving to production Hadoop deployments. It has always been focused on providing enterprise-ready, production Hadoop for analytics and transactional workloads. To do that it replaced some open-source components with proprietary pieces including its Direct Access NFS for real-time data access. It does support open source Apache projects such as Apache Drill and maintains 100% API compatibility with Apache Hadoop.

Users report that MapR’s Hadoop distributions provide high performance and reliability, and it now has more than 500 customers. Its challenge is to differentiate itself from its larger rivals and gain mindshare in a noisy market. And it needs to keep its technology ahead of the Hadoop cutting edge blade.

The pure-play Hadoop vendors, Kelly concludes, will all be looking at possible initial public offerings or acquisition in the next 12-18 months. They also may be buffeted by larger market forces that include the impact of AWS and other cloud players on new Hadoop deployments and how aggressively large players like IBM and Pivotal will mark down their Hadoop products to sell their higher-margin database products.

As with all Wikibon research, Jeff Kelly’s complete Professional Alert is available on the Wikibon Web site without charge. IT professionals are invited to sign up for free membership in the Wikibon community. This allows them to participate in Wikibon research and other events and to publish their own questions, comments, Alerts and white papers on the site.