By: Craig Wentworth, Principle Analyst, MWD Advisors
Published: 24th July 2014
This work is licensed under a Creative Commons License
My colleague Helena Schwenk and I have been tracking movement in the fast-changing Big Data management market.
As we’ve said in our recent vendor landscape report and webinar, big name vendors have tended to acquire, partner, and tweak their way into constructing a badged Big Data offering so they can say to their customers “look, we ‘get’ Big Data, we’ve got Hadoop… and we’ve sorted out the plumbing so you don’t need to worry about a thing”.
Teradata’s story so far has followed this trajectory. Three years ago it acquired Aster Data, whose product for online analytic processing was its first foray into MapReduce (and came with a SQL framework as well). A couple of years later, Teradata’s Aster Discovery Platform started supporting Hortonworks Hadoop nodes and it looked like the company’s Big Data bets were on an Integrated Data Warehouse / Aster Discovery duet.
Last year, though, it announced a closer partnership with Hortonworks—which is now supplying its Apache distribution as the basis of Teradata’s own Hadoop Portfolio (which now sits beside the Data Warehouse / Aster components as part of its Unified Data Architecture). Support for Aster continued, but the position of Hadoop did seem to be shifting (though Aster was still the place to get your SQL layer).
Fast-forward to earlier this week, and news that Teradata has made two new acquisitions in this space: Hadapt and Revelytix. Neither comes with a large customer base, so we assume it’s more about the people and the IP. And this tells us something about where Teradata looks like it’s heading.
Hadapt (which we name-checked in our vendor landscape report) brings Teradata another analytical platform, with its own SQL-on-Hadoop framework; in Revelytix, Teradata has acquired expertise in integrating Hadoop across the enterprise—metadata management, data governance, and general ‘data wrangling’, and so on.
Now there’s been a strong story building recently about what Hadoop needs to do / have in order for it to get the keys to the enterprise. These customers expect enterprise-grade performance, availability and security and so now many data management vendors (Teradata included) tweak their Hadoop offerings to make them more palatable in those respects—sometimes it’s playing with the filesystem that does it; sometimes it’s integrating a more robust security layer, etc.
Then there’s the need for Hadoop to play as part of an integrated suite (to aid with data enrichment from other sources, and the ability to feed over-arching analytic platforms)—and stop any silos hardening around MapReduce. Well, Revelytix’s productivity tools look like it could help there (so another tick for Teradata, assuming they can make it work right across their existing ecosystem—which already includes data warehouse and discovery components alongside Hadoop).
Another key Hadoop mainstreaming factor is SQL. Enterprise customers typically have large SQL talent pools (if they have any history of crunching data warehouses), so the provision of standard SQL ways in which to query Hadoop data is being touted as the serious Hadoop distribution’s must-have feature as it crosses into regular IT data management architectures. Oracle recently announced a Big Data SQL option for its Big Data Appliance—designed to run a single SQL query against Oracle database, Hadoop and NoSQL stores. Now, you might have thought that Aster already brought Teradata some of that capability… well, it seems, not enough—hence the acquisition of Hadapt.
The other noteworthy element here is that Teradata appears to be planning a steady and, one might say, cautious approach to integrating its new assets. According to reports the company is combining the employees and intellectual property of both acquisitions within Teradata Labs, the company’s research and development arm rather than fast tracking the process of bringing new offerings to market. This would seem to suggest that Teradata is also keen to utilise the engineering experience and expertise of these start up companies, as well as their IP, within its own product line, rather than just reaching out to their existing audiences which are predominantly Hadoop-based communities.
What the company does next will be interesting, though: will Hadapt’s technology be folded in to bolster analytics and SQL capabilities under the Aster Discovery platform… or will we start to see a stronger emphasis on its flagship Teradata platform and how it builds out Hadoop from its Hortonworks offering? If the latter is true then the positioning of the Aster Discovery platform is perhaps a little unclear. Furthermore, what does all this means for Teradata’s Unified Data Architecture? Teradata has yet to brief analysts on its latest acquisitions, but as soon as we know more we’ll let you know.
We automatically stop accepting comments 180 days after a post is published. If you would like to know more about this subject, please contact us and we'll try to help.
Published by: electronicdawn Ltd.