Why big data will be a big deal for the new HP

19.08.2015
When environmental group Conservation International (CI) needed to measure how much of the world’s rainforests have been chopped down, it turned to HP.

CI wanted to use satellite imagery to compare the amount of rainforest there today to images of the same areas three decades ago. HP engineers built a tool using the company’s DistributedR program that automatically scanned the images and categorized them, pixel-by-pixel to determine which areas were forested and which no longer are.

For an organization like CI – which collected more than $140 million in contributions last year – it is a classic example of an enterprise working with a big-name technology vendor.

For HP, it’s just the type of customer win the company needs. HP is soon set to embark on the biggest restructuring in its history when the company splits up this fall into two publicly traded companies. After doing so, the newly-created Hewlett Packard Enterprise will have a newfound focus on being not just a hardware company, but a software company too. And a big part of HP’s software push will be providing big data analytics. HP hopes to help a lot more customers like CI better use big data.

This fall HP is set to split up into HP Inc., which will house its person systems and printing divisions, and Hewlett Packard Enterprise where the company’s cash cow, its hardware division that sells servers, storage and networking equipment will live. There will be two other major components to HP Enterprise: Services (consulting and implementation) and Software.

Within that software division, big data, security and IT management will be the core pillars, says Robert Youngjohns, HP’s Executive Vice President and GM for the Software division, who noted that software makes up about 8% of company revenues, but an even higher proportion of profits. Youngjohns says HP CEO Meg Whitman is betting big that big data will help turn HP around.

HP’s big data offerings fall under the umbrella of its Haven product, which includes the company’s Hadoop distribution (which it works with partners like Hortonworks to deliver), along with Vertica (a SQL analytics platform), IDOL (for analyzing unstructured data) and HP’s Distributed R (for large-scale predictive analytics).

“What most enterprises want is to be able to apply analytics to drive better decisions based on the data they already have,” and help manage an ever-increasing inflow of new data, Youngjohns says. HP’s big data platform, which includes both structured and unstructured database options, he argues, addresses those needs.

The products within Haven are available as software downloads or available “on-demand” as versions hosted by HP or partners. They’re not yet available as a service on other cloud platforms like Amazon Web Services, though.

Dan Vesset, vice president of business analytics research at IDC, says HP is known primarily as a hardware company in the enterprise, so it’s got an “uphill battle” to gain market share in software. HP’s still recovering from its $11.7 billion purchase of Autonomy in 2011; HP eventually wrote off $8.8 billion of value from Autonomy. Soon after the purchase Youngjohns – a veteran of Microsoft - was brought in to help reshape the company's software strategy.

Vesset says HP is going up against stalwarts in the data software market like IBM, Oracle and SAP, each of which have strong database engines and large install bases.

“(HP) has a lot of the pieces needed, and they’re going in the right direction,” Vesset said with respect to the company’s product offerings. There is still a slight bit of unknown because of the pending break-up of the company though. HP’s ability to drive software sales based on existing hardware sales is an advantage though, Vesset notes.

Perhaps an even larger threat to HP than its traditional competitors is the advent of cloud-based analytics services from vendors like Amazon Web Services, Microsoft Azure and Google Compute Engine.

For organizations like startups who already have a lot of data stored in a cloud platform, doing data analytics and processing in the cloud is a natural choice. Cloud-platforms like AWS and Google also have powerful machines that can be spun up and down as needed to crunch data. HP’s Helion cloud offering has struggled to gain market share; it wasn’t even included as a competitor to AWS, Google and Microsoft Azure in research firm Gartner’s latest IaaS Magic Quadrant report.

But the cloud doesn't work for everyone, and HP seems to be targeting customers who still want to control their own infrastructure and not port data into the cloud. CB Bohn is a senior database engineer at Etsy, the online retailer, which uses AWS for some search-related functions and some Hadoop workloads. But the company’s primary data analysis tools are managed by Etsy staff and hosted in a collocation facility. Bohn says Etsy buys commodity hardware and uses the collo facility for workloads that have constant demand for data and analytics. This is the “own the base, rent the spike” model: Cloud platforms are good for one-off jobs and unexpected high-capacity workloads. But the everyday data and processing can potentially be done more efficiently on in-house equipment tuned specifically to those workloads.

The bottom line is HP will be counting on customers like Etsy and CI to help it stave off the significant headwinds facing its business as it embarks on its restructuring journey.

(www.networkworld.com)

Brandon Butler