Failure to launch: Hadoop will be 'anaemic' for years, says Gartner
Analysts claim it'll be at least 24 months before 'something' happens
Forget the hype and the vendors chinning each other – Hadoop is still not ready for take-off.
That’s according to Gartner, whose latest short story in the Hadoop-is-not-ready saga has predicted “anaemic” adoption of Hadoop for at least the next two years.
The biggest problem is that people allegedly still can’t use Hadoop, despite recent advances in tooling. It remains something for rocket scientists, not the masses. Also, Hadoop is still too expensive for firms, which still can’t see a business reason for justifying such a commitment.
Gartner’s findings are based on a survey of 285 IT and business leaders, it said. Of these, 125 claimed to have invested in the Big Data framework or said they planned to do so in the next 24 months. But 54 per cent had no plans to follow them, while only 18 per cent said they’d commit IT budget to the Big Data framework in the next two years.
Just 26 per cent are deploying, piloting or experimenting with Hadoop, with 11 per cent planning to invest in the next year and seven per cent within 24 months.
Early adopters of Hadoop inside those firms that have adopted, meanwhile, are proving no help in driving uptake elsewhere in their organisation, Gartner said.
According to Gartner, there are fewer who plan to begin Hadoop deployments in the next two years than already have. This is unusual for a new technology when it wants to achieve break-out status or be considered a success. The essence to uptake has been winning over early adopters, who proceed to be envied, evangelise their technology or carry it with them to new departments and businesses as they move around with the job: that’s how iPads, iPhones and Salesforce have penetrated the business ranks.
Hadoop isn’t considered a priority for some of those surveyed, while for others, the problem is that Hadoop is simply overkill for the problems they are facing. This fact suggests firms either don’t have the volumes or types of data that Hadoop likes to chew up, or that the hardware and software set-up required to run Hadoop remains too expensive for the problems businesses are trying to solve.
Demonstrating the value of Hadoop is the system’s second-highest challenge Gartner, said. Forty nine per cent reckoned the problem was trying to figure out how to get some kind of value from Hadoop.
The first? That would be skills.
Fifty seven per cent called this a “major inhibitor” – something Gartner called a “key blocker". Tools vendors have been responding but Gartner reckoned what’s been coming from vendors remains only of use to “highly skilled users” and it’ll take another two to three years before this challenge is addressed.
Merv Adrian, Gartner research vice president, said in a statement: “Future demand for Hadoop looks fairly anaemic over at least the next 24 months. Moreover, the lack of near-term plans for Hadoop adoption suggests that, despite continuing enthusiasm for the Big Data phenomenon, demand for Hadoop specifically is not accelerating.”
The news is unlikely to dampen the drive from the Hadoop vendors themselves – quite the contrary.
Gartner's report was issued as the recently IPO'd Hortonworks announced first-quarter net revenue of $22.8m, up 167 percent, but a doubling of losses to $40.1m, equating to $0.97 a share. The company claims 400 customers are now paying for its support subscriptions to Hadoop.
They are fully committed to trying to make money from selling Hadoop services and subscriptions, and convincing you that you really must use the system. Gartner reckons their best hope, for now, is moving into even larger deployments within existing customers.
The distro vendors claim customers in banking, telecoms and retail, in areas including fraud detection, customer churn and services, and buying analysis. On Wednesday, Cloudera named UK retail giant Marks & Spencer as a sign-up to Cloudera Enterprise, Data Hub Edition, to help on data analytics.
It’s a step up from the IT department, where Hadoop had gained its footing and seemed destined to stay performing bread and butter log crunching work. ®