We had the opportunity to attend the Teradata Influencer Summit held at the beautiful L’Auberge Del Mar in northern San Diego. When mentioning to other influencers that I would be en route for that meeting earlier in the week, I mostly gathered incredulous stares and comments like ‘are they still around and interesting’? I missed the first half day, but the next one and a half days gave a good insight into where Teradata is and where they want to be in the next years. 

 
 
 

So here are my top 3 takeaways from the event

A compelling vision – In recent times Teradata had been blamed for a lack of vision and / or a lack of sense of realism. I followed the Summit two years ago from the fences and back then an executive had an elaborate presentation to show why Hadoop would be a failure – or only a temporary fashion that would disappear as fast as it showed up. But quickly afterwards (and some colleagues say the Summit helped) Teradata reversed course, changed from a ‘struggle to snuggle’ approach vis a vis Hadoop, embracing more of Hadoop, partnering e.g. with MongoDB already a year ago (and again was a key sponsor of this year’s MongoDB World – also this week, event report here). All slides at the summit had Teradata, Aster and Hadoop based offerings. As a matter of fact the Teradata UDA includes all three database paradigms.
 
The vision of the 'Sentient' Enterprise
And along these capabilities Teradata is using its by now proven Querygrid capabilities, following the ‘co-existence’ route with Hadoop, or ‘federated query’ strategy all vendors with an age of 5+ years pursue. As part of that Teradata sees the world separated into three categories of systems – tightly coupled (Teradata), loosely coupled (Aster) and non-coupled (Hadoop). An interesting approach to the system landscape, that Ratzesberger used to explain the vision of the ‘sentient enterprise’. There are pros and cons with federation, see MyPOV section below.


It’s coming together – Not sure when Teradata came up with the vision, but obviously work needs to be done here. Teradata needs to write the endpoint to execute queries on the Hadoop side, and it currently supports Hive for that. But Hive has known benefits and challenges, and Teradata will address the challenges soon, stay tuned on that topic. The other piece critical is that a tool that can transport data back and forward, extract as needed from a variety of data sources as desired, transform information, can apply rules ‘in flight’ etc. Teradata calls it a Listener, but it is much more. In the old times one could call it a bona fide ETL tool, only that today it scales to the 21st century requirements with streaming data, humungous data volumes, BAM and CEP like support etc. – a key product / tool to make the Teradata UDA fly. 
 
 
 
The Analytical Application Platform

Self-disruption built in – All successful technology companies struggle with disrupting themselves. It’s a price of success and harvesting the benefits of maturation of technology that create the potential for technology disruption. Co-President Wimmer remarked that all successful database companies, the 30+ year olds are setup for disruption today. Usually technology players react to disruption risks with acquisitions, try to insert knew product capabilities and employee talent and these approaches are usually less successful than intended for a variety of reason.

Teradata has taken a unique approach here, allowing the acquisition of ThinkBigA to keep operating independently, and even taking a step further, funding its global expansion. Unique, because ThinkBigA does pure best of breed BigData projects, with complete independence of choosing technologies customer’s desire. As such ThinkBigA cannibalizes Teradata customers to a certain extent, and remarkably Teradata does nothing to stop that.
 
The Teradata free ThinkBigA architecture
So kudos to the Teradata management, that coming of a close to death experience in regards of the Hadoop phenomena, now has a built in disruptor with ThinkBigA. Quite a gutsy move, coming close to e.g. an enterprise ERP vendor allowing a subsidiary services company to sell services on competitor products to its customers. What? Granted the ThinkBigA case is not that drastic, as the provider ‘only’ uses open source, but the disruption is on hand. The benefits for Teradata are twofold: The vendor learns about real customer demands and quality of open source BigData products, that it can use for its own R&D strategy and effort. Secondly Teradata has a way to remain relevant (and extract revenues) from customers it otherwise would have lost to BigData / Opensource players anyway.
 
ThinkBigA BigData Project recommendations, notice bullet #5 - emphasis added
 

MyPOV

Teradata is alive and well, with solid product plans for the immediate future. The three tiered vision of federated queries and data is common for vendors with existing, substantial business. It resonates well with enterprises that are more conservative both in regards of technology adoption and security needs.

On the concern side, the verdict is out if enterprises operating with a federated storage and query approach will end up missing insights, or only getting served sub optimal insights. Conceptually one can make the case that every insight can only be caught with a single source storage, one place to be queried etc. but if the difference between the two is substantial for the success of a business, remains to be seen.

So a key 2nd half for Teradata on the product side, with a number of substantial announcements to come, key capabilities in the pipeline. Teradata may have been down, but it’s back on its feet and throwing punches. Only the paranoid survive and Teradata has gotten a litte more paranoid - in a positive way. Good to see for prospects, customers and the overall ecosystem as nothing fuels innovation better than competition. We will be watching.
2012, 2013, 2014 & 2015 (C) Holger Mueller - All Rights Reserved

Business Research Themes

Data to Decisions

Roles

Chief Information Officer

Tags

analytics, hadoop, nextgen apps, Teradata