Sorting the Hadoop Opportunity
There’s no doubt Hadoop is one of the more disruptive enterprise technologies to infiltrate the enterprise in quite a long time. The question many solution providers are wrestling with now is how much opportunity that disruption creates for them.
There’s no doubt Hadoop is one of the more disruptive enterprise technologies to infiltrate the enterprise in quite a long time. The question many solution providers are wrestling with now is how much opportunity that disruption creates for them.
At its base level Hadoop is being heralded as a “data lake” for inexpensively collecting raw data that can be processed more efficiently elsewhere. To a certain extent that’s true. But it’s also becoming apparent Hadoop is a platform in its own right where multiple types of applications can now be processed. In that context, Cloudera CEO Tom Reilly said Hadoop is more of a hub for processing data both in and out of the cloud.
Vendors, he said, are aligning their Hadoop efforts around those two visions of Hadoop. Cloudera partners such as Oracle (ORCL), Intel (INTC) and, most recently Teradata, view Hadoop as a platform that fundamentally will alter how data warehousing applications are processed, Reilly noted. As such, Cloudera is investing in developing extensions to Hadoop that will enable such an occurrence.
Microsoft (MSFT), Hewlett-Packard (HPQ), SAP (SAP) and Red Hat (RHT) have pretty much decided to back Hortonworks as their preferred distribution of Hadoop. Each of these vendors are adding their own extensions to Hadoop to create next-generation data warehousing applications. In the case of Microsoft, Eron Kelly, general manager for the Microsoft Data Platform, noted within Microsoft Hadoop is seen as both a lake from which to draw data and ultimately a platform for processing big data analytics. It also means solution providers such as Avanade, a joint venture between Accenture and Microsoft, are on the Hortonworks bandwagon.
Most solution providers probably would be better off developing their Hadoop strategy in the context of the maturity of their customers. Many customers have extensive investments in data warehouses that they can’t afford to replace overnight. In those instances, Hadoop for now is primarily a source of new and very interesting data. There are, however, many organizations that consider big data to be enough of a strategic imperative that they have begun building big data applications directly on top of Hadoop. There may not be as many of these customers out there just yet, but among the one there are the investments being made in these applications will quickly approach millions of dollars.
Against that backdrop, solution providers should be making investments in acquiring big data skills today. At the rate it’s disrupting traditional IT thinking in the enterprise, it’s more than likely Hadoop, in one form or another, will go into production in more than one customer environment sometime in 2015. And once it does, nothing about how that customer processes, manages or stores data will be the same again.