MapR Reference Architecture Makes Big Data Practical

The world of Big Data continues to mature, a fact made clear this week with the release of a formal reference architecture for Big Data developed by MapR and Hewlett-Packard (NYSE: HPQ). The architecture, described in a white paper released on HP’s website, promises to help make it easier for organizations to conceptualize, plan and deploy the software and hardware components needed to build next-generation data infrastructure.

MapR, a relatively young and privately held company that distributes value-added implementations of the open source Apache Hadoop platform for Big Data, has forged a number of partnerships in recent months with enterprises such as Talend and Hadapt. It has also expanded the geographic scope of its operations to Europe.

Now, MapR’s engagement with the channel has expanded to include a major hardware vendor in the form of HP. Together, the two companies wrote a reference architecture that they envision “as a blueprint for their success when working with Big Data datasets,” according to a statement. The move should not only strengthen HP’s role on the Big Data scene, but also help to promote tighter integration between the hardware end and the software end of Big Data.

Creating a Big Data Language

The story goes further than that, however. Perhaps the most important takeaway point from this news is what it says about the emergence of a new vocabulary for thinking about Big Data in hard, real-world terms.

Like other buzzwords of the IT lexicon–“the cloud” being a key example–Big Data is a term that’s easy to throw around. But understanding what it actually means in a specific, practical sense can be harder. Reference architectures and similar documents help to solve that conundrum by bridging the divide between the abstract and the practical. They make it possible to speak very precisely, with specific examples, about the hardware and software solutions an organization needs to bring together to develop a Big Data strategy.

The reference architecture released this week by MapR and HP, of course, privileges the products and particular viewpoints of those two companies. It’s not a totally comprehensive frame of reference for the Big Data world as a whole–although the open source core of MapR and the ubiquity of Hadoop mean that the software component of the reference architecture will have particularly wide applicability.

And in any case, the document is a crucial starting point for addressing Big Data issues in a hands-on, real-world way. It also shows that Big Data is transitioning from a hot discussion topic among techies into a standard, easily implemented resource for a wide variety of organizations.