The folks at HP must have gotten the memos about trends such as NoSQL and the Internet of Things, and they decided to roll the requisite capabilities into the seventh edition of the company’s Vertica database called “Crane.” Once one of a handful on columnar databases on the market designed for fast analysis of relational data using SQL, the latest edition of Vertica supports the analysis of unstructured data at a price point more in line with the often free, open source price associated with many NoSQL databases.
To be clear, though, Vertica did not turn itself into a NoSQL database, Luis Maldonado, director of product management at HP Vertica, explained in a recent interview. Rather, it’s now sporting a product called FlexZone that integrates with the Vertica platform and “allows [users] to connect to any arbitrary source” via simple connectors to and pull data (e.g., JSON files from MongoDB and CSV log files from Hadoop or someplace else), he said. Everything comes together in a single view of the data, and there’s no need to do lots of work on the data before analyzing it using tools like Tableau or QlikView, because Crane has a capability called “auto-schematization.”
FlexZone, however, is only an entry-level product, which Maldonado says is “disruptively priced.” He described it as bridging the gap between leaving data sitting dark in vast lakes like Hadoop and giving it the full-on data warehouse treatment like the enterprise version of something like Vertica provides. He acknowledges that Hadoop — especially as that platform takes on more Vertica-like performance characteristics and SQL capabilities — actually inspired the FlexZone price point (and is generally driving prices down from Vertica’s one-time price of $100,000 per terabyte) by making users think twice about where they want to store data.
FlexZone users who want to upgrade their data to the enterprise version and get the performance and other capabilities that come with that can do so in a single step.
There’s something of a great middling going on in the data management world right now, and it’s not clear exactly how it will play out because no one platform is yet good enough at everything to actually replace the other stuff full stop. Hadoop is for real and every database and data warehouse vendor is bending over backward to integrate with it, even while it’s increasingly moving into their realm with more enterprise features and SQL capabilities. On the other hand, database vendors, including Vertica rival Teradata, are moving in on some NoSQL databases by adding support for JSON files (which are MongoDB’s sweet spot) and key-value capabilities.
Right now, it seems like it’s still largely a matter of getting what you pay for in terms of capabilities — even if that means paying for (or at least supporting) five different systems. Maybe that will change over time and the data layer will flatten out in one system that can do it all, or maybe the future of data is just more complex (and possibly more expensive) than ever.
Correction: This post was corrected at 11:38 a.m. to properly identify FlexZone as a separate product, not a Vertica feature.