As Hewlett-Packard Co.’s Big Data Conference 2015 kicks off this week in Boston, customers will be looking for evidence that the hardware giant is evolving its strategy in software and can be a strategic supplier for Big Data and analytics initiatives. To do so, say Wikibon analysts, it needs to fill gaps in the platform it has stitched together with a series of piecemeal acquisitions and investments, and deliver on the need for a new generation of analytic platforms that the analyst community calls Systems of Intelligence.

At the core of HP’s big data strategy is Vertica, which along with other massively parallel processing (MPP) database suppliers ushered in a new generation of database systems last decade. MPP platforms aren’t a replacement for traditional relational DBMS, said Wikibon analyst George Gilbert. Rather, they’re an evolution that accommodates the growing need to handle unstructured and streaming data in a way that enables faster decision-making and predictive analytics. “Unlike traditional Systems of Record such as SAP, the master data about a customer – the customer profile – evolves and grows continually,” Gilbert wrote. “The traditional SQL DBMS has had great difficulty accommodating that.”

HP brings a mix of technologies to the table to attack the emerging data analytics market. With its 2011 acquisition of Vertica, the analytic database that was the springboard for this conference, HP got best-of-breed technology for analyzing large amounts of data using the tried and true SQL language. Gilbert called Vertica, “the best column store DB on the market.”

The 2011 acquisition of Autonomy Corporation PLC was a disaster from a business perspective, but did bring some powerful technology for enterprise search knowledge management and governance, all useful elements for building Systems of Intelligence.

HP also has a $50 million investment in Hortonworks, Inc. securing it a seat at the table in the Hadoop market, which is always a major topic at Big Data conferences. HP is also an acknowledged leader in scale-out hardware and enterprise security, two other critical factors for building out Big Data platforms.

Haven connection

The company’s software division ties these disparate technology pieces together under the brand name Haven, which HP calls “the industry’s first comprehensive, scalable, open, and secure platform for Big Data analytics.” But Wikibon analysts say HP needs to evolve its developer ecosystem to fill in the gaps and be able to scale.

An analytics database is only part of the picture, Gilbert said. Systems of Intelligence is a pipeline concept in which discrete components for tasks such as streaming and graph processing can be plugged into the data flow depending upon the application. For example, users may need to use streaming data processing, applied machine learning or graph processing on data sets in the quest to anticipate customer needs. “Vertica works really well as a high-performance engine when you’re starting out,” Gilbert said. “I see more challenges as it moves into the world of mix-and-match engines.”

“Having the best MPP is good when you’re offloading workload from a traditional warehouse to Hadoop at much lower cost, but you need other products to build that pipeline, and HP doesn’t really have all the pieces,” Gilbert said.

Then there’s NoSQL, the highly scalable clustered database technology that is increasingly challenging traditional relational databases for analytical and even some transactional applications. “SQL is being supplanted by NoSQL,” summed up Wikibon co-founder David Floyer. While Vertica is considered one of the best SQL engines in the Big Data arena, HP doesn’t have a strong NoSQL story to tell outside of some partnerships. It also lacks a strong story in graph and streaming engines.

Use the source, Luke

Given that many of the innovations in these areas are springing from the open-source community, HP would probably do best to concentrate its energies there, Wikibon analysts agreed. For example, a new breed of tools like Flink and Spark, both developed in the Apache ecosystem, may effectively fill the real-time gap in HP’s portfolio. But the company has to redouble its efforts in open-source software development. Wikibon says that HP’s contributions to the OpenStack community should serve as a good model for its Big Data efforts.

While HP can leverage partner and open source technologies to “plug” these pieces into Haven, Wikibon analysts said HP must accelerate its efforts to build a more vibrant developer community to truly achieve scale. HP has not been known for having a deep and wide community of developers and, in part, its Big Data Conference is trying to address this shortcoming. “IBM has an open source community in its DNA,” said Floyer, drawing a comparison. “HP still doesn’t have it.” Nor does it have a strong record of building and supporting developer communities, he added.

All this doesn’t mean for a minute that HP won’t be a major player in Big Data and Systems of Intelligence. In addition to having many excellent point products, it has a strong services organization that understands how to make all the pieces fit together and work. HP also has one other characteristic that most of its enterprise brethren lack: a tolerance for lower-margin businesses.

HP has stuck it out and learned to make a profit in businesses like servers and workstations where competitors have bailed, noted Wikibon Chief Analyst David Vellante. “HP has gotten comfortable with the lower-margin economics of giving customers more choice,” he said.

In an enterprise IT world that is being convulsed by crumbling profit margins in the face of white box and open-source competition, that survivalist culture may turn out to be HP’s strongest asset.

This interview with Alan Nance, VP of technology at Koninklijke Philips N.V. (Royal Philips) at the 2014 HP Big Data Conference was judged one of the best theCUBE videos of the year by viewers. In it, Nance describes how Royal Philips is overhauling its entire IT architecture around services and cloud-like payment models (28:54).