Teradata Broadens the Base

Del_Mar_BeachIn recent years, many were thinking that Teradata was losing its edge in the greater analytics marketplace. Long a favorite for large enterprise customers seeking to do world-class data warehouses, the company realized the market was rapidly shifting around it. Two major trends seem to have the biggest impact here: big data and cloud. While Teradata's core install base isn't at significant risk of going away, they may be looking at complementary alternatives. For the company to remain a dominant player, some new approaches would be required. The recent Teradata third party influencers event in Del Mar, California, showcased just how far their efforts at transformation have progressed.

First up is cloud-based analytics and data warehousing. Described as "more theory last year," now real customers are live in 2 data centers. The major goal is to expand footprint in existing accounts, bringing appliances and cloud together, though with cloud more aimed at departmental, DR, or test/dev use cases. At the same time, it's clear the company would like to win new converts, not least by making it more accessible to midmarket customers that may otherwise be priced out of using Teradata. Of course, there is always a tipping point where on-premises becomes more cost-effective again.

In developing the Teradata cloud, the design focus was on high performance, making it multi-tenant and secure, and offering relatively fast 24-48 hour provisioning. Well, fast compared to having an appliance shipped, installed, and configured anyway, if somewhat slow compared to other on-demand cloud services. Advanced services include: BI and ETL server provisioning, more networking, more security, and of course lots of consulting to identify, advise, architect, implement, optimize, and manage the use cases.

The second big development is around big data environments. Clearly, traditional data warehousing isn't going to be as economical as Hadoop for large volumes, yet it can offer other advantages, like analytics performance and maturity. Teradata is both embracing Hadoop and striving to clarify (their view) of where it fits best. Teradata argues that Hadoop does have many limitations today and much of the total cost of integration and operation is hidden behind promises of open source software and commodity hardware. A healthy coexistence seems likely for most customers. 

Teradata detailed plans to get more squarely connected in the Hadoop ecosystem, some imminent but not yet announced (I won't spoil their news!) The general themes were around how to integrate across businesses for related data types and analytics, especially where flexible querying, supporting real-time and streaming analytics, effectively managing and governing data lakes, avoidance of (expensive) ETL/ELT, and multi-functional models may be required. It was quite clear that Teradata recognizes the market movement to Hadoop and wants to bring it in alongside their traditional offerings. Whether this reads as embracing Hadoop or being defensive depends on your own personal level of cynicism, but either way it is the right move for the company.

Also worth noting is that Teradata has two not-so-secret weapons to enhance their traditional product strengths. One piece is Aster for fast and powerful advanced analytics, with 120 engines to handle all sorts of specific needs, such as statistical models, machine learning, behavioral, text, or graph analytics. The other is the worldwide professional services and consulting expertise to make this stuff all work in complex customer environments, and the Think Big acquisition has been given room to operate freely and semi-autonomously in big data areas. In all, I came away with renewed respect for Teradata's vision, and more importantly its ability to re-invent itself and execute against that vision.

big data analysis


Topics: Data Platforms, Analytics, & AI