I recently attended Teradata’s Third-Party Influencers Meeting (Twitter hashtag #TD3PI) in Las Vegas where the company updated industry analysts and consultants on upcoming product plans and the status of two product lines it added via acquisition in the past year.
Randy Lea, Teradata’s VP of product management and marketing, provided a company overview recapping highlights and defining the corporate strategy building on its work in 2010 that my colleague assessed. Teradata focuses on three markets: its core business of data warehousing, which it identified as a $27 billion market, the $15 billion business applications market represented by its Aprimo acquisition, and the big-data analytics market, a $2 billion market addressed in part by its Aster Data acquisition. I mention Teradata’s estimates of market size as they may indicate the emphasis and investment the company will make in each segment.
Lea referred to Teradata’s independence from NCR, for almost four years, as a good thing. Independence, he said, has allowed Teradata to focus, to release products more frequently and to acquire companies that enhance their market position. Judging by the performance of its stock (NYSE: TDC), the financial markets seem to agree independence has been good – even with the recent market downturn its stock is up 75% in the last year.
Over the last few years the Teradata product line has evolved to include a family of systems. The product line was succinctly summarized as follows:
- the 6600 line for data warehousing, the bulk of Teradata’s revenue and its core market
- the 2000 appliances for data marts
- the 1650 “extreme data” appliances for deep history
- Aster Data for MapReduce
As part of the meeting Teradata previewed some of the features in its upcoming release Teradata 14. Perhaps most significantly, Teradata will be adding columnar database capabilities to its product. These features will provide better performance and better compression for many types of common analytical workloads. (See my previous post for more background on columnar database capabilities.) We’ll have to see how well it performs compared to other columnar implementations, but from the company’s presentation it appears to be a robust implementation of columnar capabilities. Teradata customers can benefit from taking advantage of columnar processing without having to adopt another technology platform for that purpose and can manage it all from within the Teradata environment.
As the product family above suggests, Teradata has embraced the notion of an analytical ecosystem consisting of multiple systems. In some cases multiple systems provide fail-over; in other cases the additional systems are for specialized workloads. Teradata 14 includes features to make it easier to leverage and manage these multiple system environments. Specifically, new query management capabilities will provide intelligent query routing and data synchronization across these systems.
Earlier this year Teradata introduced its Active Data Warehouse 6680, which includes a hybrid solid-state drive (SSD)/hard disk drive (HDD) configuration. The hybrid configuration leverages Teradata Virtual Storage (TVS) for managing and optimizing data across different types of storage devices. Since data can be retrieved more quickly from SSD than from HDD, it makes sense to try to put the most frequently accessed (or “hot”) data on SSD and the less frequently accessed (or “cold”) data on HDD.
Teradata aims to make everything automatic. No user configurations or settings are required. Every five minutes, TVS moves block-level data between devices based on usage statistics. So on Monday morning, for instance, when most retail operations are evaluating the prior week’s sales, that data will move to the better-performing drives. Similarly at the end of a quarter or year as the finance department closes the books and prepares various financial reports, the associated data will move onto the SSDs. Over the next few years, while the price differential between SSD and HDD is still significant, organizations are not likely to purchase 100% SSD systems. Hot and cold data solutions such as TVS can help customers squeeze the most out of the SSD they do buy.
Tasso Argyros, co-founder of Aster Data, announced that Aster Data will continue to operate independently. Teradata positions Aster Data as a solution for “multistructured data” and advanced analyses, with a heavy emphasis on SQL-MR, its proprietary, patented MapReduce implementation. As I suggested in my blog post on the acquisition, Teradata seems to want to compete with Hadoop via Aster Data’s MapReduce capabilities.
While they may be able to handle similar workloads, SQL-MR cannot participate in the same community of products, tools and expertise that surrounds Hadoop MapReduce. Argyros did share a few details under nondisclosure about technology investments that have been made since the acquisition that it will be announce in the future. And although the acquisition was only completed in April, I would have liked to see a more complete vision of how Aster Data evolves in the long run within the Teradata product family. Randy Lea announced that he will be changing roles to lead North American sales for Aster Data. Moving a key executive into the new division suggests that Teradata is committed to Aster Data. Lea’s direction could also help drive further integration of the organization and its products.
Teradata has consolidated its marketing applications including Teradata Relationship Manager in the Aprimo division. Lisa Arthur, CMO of Aprimo, gave an overview of the Integrated Marketing Management product line and progress that has been made since the acquisition. Marketing analytics are common within the Teradata customer base, so expanding into marketing applications is not a big stretch. However, to be a serious competitor in the business applications space, Teradata will need to build a broader portfolio of applications. There was little discussion about how, when and whether Teradata intends to create such a portfolio, but I would not be surprised to see additional acquisitions.
Overall, Teradata appears to be executing well; I would also characterize the company as maturing as an independent software vendor. Its products are generally well regarded, although perhaps overpriced, and it has shown a willingness to invest and expand beyond the original high-end product line. With the appliance product family it now offers different entry points for new customers. The scenario still has a few areas that need to be developed further, but if you are considering database technology to support analytics in your organization, it’s worth exploring what Teradata has to offer.