Wikibon Blog

With all the talk of converged infrastructures and the journey to the cloud, data analytics has gotten short shrift at HP Discover this week.

That’s understandable if you consider HP’s marketing strategy. The Palo Alto-based technology giant is betting the future on its ability to deliver integrated server, storage and networking solutions and is determined to draw a distinction between what it considers its open standards approach and that of competitors like Cisco and Dell.

But a converged infrastructure — any infrastructure really — is only as good as the applications it supports. Among the most popular at the moment are applications that perform analytics on ever growing volumes of structured and unstructured data – Big Data Analytics if you like. For HP, Big Data Analytics means Vertica Systems, the columnar-oriented, massively parallel processing (MPP) data warehouse vendor it acquired earlier this year.

(SiliconANGLE Founder John Furrier and Roku’s Brian Jacquet discusses HP’s acquisition of Vertica Systems.)

Vertica is likely to play an increasingly important role in HP’s converged infrastructure push. We got just a taste of what’s to come today with the release of HP Vertica Analytics System. The new analytic appliance includes Vertica’s data warehouse preconfigured to run on HP’s converged infrastructure.

Specifically, the new appliance — available in quarter-rack, half-rack, and full-rack configurations — is built on HP BladeSystem server blades and HP StorageWorks. HP claims it can be optimized to store and process up to a petabyte of data on a single rack. It uses Linux as its operating system and comes preassembled for quick deployment.

Even before HP acquired it, Vertica racked up a number of impressive customer wins. Among them was Zynga, the social gaming company responsible for Facebook hits Farmville and Mafia Wars. Zynga began using Vertica a little over a year ago as it struggled to analyze huge volumes of user data. At the time, the company supported over 40 million daily active users generating 3TB of new data a day, according to Ken Rudin, Zynga’s Vice President of Analytics.

Zynga tapped Vertica to run real-time analytics against all that user data to determine usage trends and to test new game features. According to Rudin, Zynga’s implementation included 230 nodes spread across two clusters of Vertica’s data warehouse.

Part of what sets Vertica apart from traditional data warehouses is its data compression capabilities – it offers a dozen unique compression schemas — and its hybrid in-memory/disk-based architecture. The result is fast data loading and query times, like that at Zynga, without overtaxing the server and storage hardware.

Vertica was also designed from the ground-up for cloud deployments, and was among the first data warehouses deployed in Amazon’s EC2. It also added support for and integration with Cloudera’s commercial Hadoop distribution way back in 2009, which is ages ago in Big Data terms. Both make Vertica well suited for converged infrastructures like that offered by HP.

HP is hardly alone in the Big Data analytics space, of course, and faces fierce competition from rivals ParAccel, Aster Data and Greenplum (the latter two recently acquired by Teradata and EMC, respectively.) HP may have a leg up, however, due to the breadth of its hardware offerings and the increasing popularity of the appliance delivery model. Expect to hear Vertica mentioned with increasing frequency in HP’s marketing message as the company furthers integrates the data warehouse technology with its converged infrastructure offerings.

HP shops should strongly consider Vertica for Big Data Analytics workloads, especially those with heavy read-intensive workloads and/or extensive Hadoop deployments. Non-HP shops and those with heterogeneous data center environments should also include Vertica in their Big Data Analytics evaluation process, but be sure to press HP on its commitment to support Vertica deployments powered by non-HP hardware.