Projected layoffs, disappointing earnings and stock price dips have befallen companies that offer Hadoop-based analytics solutions geared at big data in the past two weeks.
Toward the end of May, Hadoop-based big data company MapR filed a WARN notice with California labor authorities. Citing "extremely poor" results on its quarterly earnings report, last-minute client pullbacks and the inability to find new funding, it expected to let go 122 staffers at the company's Santa Ana, Calif. headquarters by June 14.
Last week, MapR competitor Cloudera — a clear front-runner in the field following its $5.2 merger with Hortonworks last year — shed a third of its value after its stock price took a nosedive.
The trigger for the event: the departure of CEO Tom Reilly and the release of disappointing earnings projections.
It’s a surprising outcome for Hadoop, an open-source framework launched by Apache in 2006, initially lauded as a key component for the storage and processing of big data. Now, the spread of cloud-based platforms like Amazon Web Services or Microsoft Azure has been luring customers as they focus on real-time data and event-driven applications.
As technology evolved around it, the decline of Hadoop felt unavoidable. Industry observer Forrester predicted in a 2017 report, The Cloudy Future of Hadoop, that the technology would face challenges from larger cloud players, with artificial intelligence applications eclipsing its reach.
"That is exactly what has happened," said Mike Gualtieri, VP and principal analyst at Forrester Research, in an email to CIO Dive.
Companies like Microsoft or Amazon allow customers to deploy Hadoop and related technologies in the cloud, and artificial intelligence became the biggest appeal for large data analytics projects. MapR failed to respond strongly enough to either trend, said Gualtieri.
Ashish Thusoo, co-founder and CEO at Qubole, said migration of big data storage from Hadoop-based file systems to cloud object stores provided by AWS, Azure, and Google eliminated the need for on-premise HDFS file systems like the one provided by MapR.
In addition, there's a preference for platforms that can process data right where it is, Thusoo said.
That said, Hadoop is an umbrella term for a range of technologies which continue to remain relevant, the executive said.
"Hadoop continues to serve the needs of organizations that have successfully adopted it on-premises," said Thusoo. "What we are witnessing now is an evolution of this technology for different deployment platforms (cloud) and different use cases."