Reaction Time: Teradata Users Express Cautious Optimism
What do Teradata users think about the data warehousing specialist’s upcoming split from parent company NCR?
- By Stephen Swoyer
- February 14, 2007
We know what analysts and Teradata officials say about the data warehousing giant’s upcoming split from parent company NCR Corp., but what about Teradata users? They’re the ones, after all, who have the most to gain—or lose—by the upcoming division; so what’s their take on things?
Based on discussions with Teradata users, it seems that many are cautiously optimistic about the looming split. They seem to like the idea of an unfettered, unlimbered Teradata: one that’s free to concentrate solely on the bread-and-butter data warehousing segment; and free, perhaps, to explore new opportunities on the data warehousing low-end, as well as in the mid-market. However, some Teradata users seem to be worried about Teradata’s viability as a separate entity.
"I was not surprised to hear this," says Teradata administrator Jeff Ohlman, who asked that his company not be named. "I think the jury is still out as to whether this will benefit Teradata. Obviously it will allow NCR to focus on its core business, but I think the danger of Teradata being destroyed in a buy-out is high."
On the plus side, Ohlman says, an independent Teradata will be able to focus solely on evolving and enhancing its relational database, warehouse, and universe of associated tools. "NCR has neglected Teradata for years. As a result, the functionality in comparison to other RDBMSes is way behind," he argues, citing two among several relevant pain points: "[Teradata’s support for] stored procedures and [its] user interfaces are weak at best."
Fred Pluebell, a Teradata administrator with a prominent brick-and-mortar retailer, expresses optimism about the impending split. "On balance, I think it's moderately positive. Under Bill Nuti, Teradata seems to have been operating more independently of the rest of NCR already, so I don't think it will mean big changes right away," Pluebell comments. "But it does give [Teradata] the opportunity to raise capital in the marketplace with [its] status as a leading technology company, should they have the need."
Pluebell says he was never persuaded by the notion of "synergy" between two such dissimilar companies as NCR and Teradata. In this sense, he suggests, the split is long overdue. "[N]one of Nuti's statements ever seemed to express the ‘synergy’ between NCR divisions that Lars Nyberg and to a lesser extent Mark Hurd used to talk about—but could never demonstrate. So I think if Teradata were a short-term takeover target, NCR would have already sold them."
Teradata administrator Nolan Madson, who likewise requests that his organization not be identified, says the looming NCR/Teradata split is "a good thing." Madson says he doesn’t feel NCR has been a bad corporate parent to Teradata—not by a long shot—but he does note that "Teradata hasn’t been [NCR’s] only focus." That will change when Teradata becomes its own boss, Madson says, and it’s possible Teradata’s new leadership could introduce other changes, too—especially on the pricing front. "[M]y organization is running into real budget problems, and the expense of Teradata storage is consuming time that could be better spent on solving business problems," Madson comments.
Starting in 2002, Teradata locked horns with Netezza Inc.—a then-little-known data warehousing appliance vendor—in the mid-to-high-end DW segment. Things got even more fractious in 2005, when a second appliance vendor (DATAllegro Corp.), as well as a bona-fide computing giant (IBM Corp.), jumped into the DW appliance fray. Since then, another couple startups (Calpont and Dataupia) have joined the battle, while both Sun Microsystems Inc. and Hewlett-Packard Co. (HP) have thrown their not-insignificant hats into the ring.
Teradata, for its part, has pursued a mixed strategy for some time now, essentially saying, "We’re an appliance but we’re not really an appliance." The company touts its bundled hardware and massively parallel processing (MPP) architecture as appliance-like value-adds; but also cites its custom RDBMS and highly optimized data warehouse (along with a proprietary NCR Unix powerplant that’s also tweaked for the Teradata RDBMS and Teradata Warehouse) as competitive, non-appliance-like differentiators.
As a result, users like Madson say that Teradata will consider fielding a line of "white box"—or commodity—data warehousing appliances; other Teradata veterans, such as Ohlman, say they don’t see the value of a Teradata software solution that’s decoupled from its MPP hardware, however.
"I think that any chance to gain market share is a good thing. On the other hand, for the most part the only reason to use Teradata is to take advantage of its scalability as an MPP platform. [Teradata’s] functionality without MPP’s competitive advantage is… behind other systems," he argues.
Technologists who are familiar with Teradata say its MPP underpinnings comprise its most important competitive differentiator. (http://www.tdwi.org/News/display.aspx?id=8296) And now that some of its competitors are embracing off-the-shelf commodity solutions (a prominent DW appliance vendor is expected to announce such a deal with a top PC OEM sometime this month) Teradata needs to articulate a DW appliance strategy, they say.
"[Teradata] can still support larger data volumes than other databases. The problem is that current hardware market limits TD's value proposition to only the highest-end data warehouses," said data warehouse consultant (and long-time data warehousing administrator) Mark Madsen, in an interview last month. "The question is, when today I can put a terabyte of memory and a couple hundred processors to work, connected to a hundred terabytes of disk, and make a single-instance database like Oracle work, why do I want a specialized database with special knowledge needed to manage and administer it? How many companies have problems of that scale?"
Some Teradata users express concern about potential brand dilution, too. As poet A.R. Ammons once put it, "a word too much repeated/falls out of being." If Teradata isn’t careful to differentiate its own discrete appliance offering from those of its competitors, there’s a chance it could in turn dilute—that is, compromise—the impact of its high-end data warehousing brand.
"From a purely technical point of view, Teradata has certainly been capable of running on other hardware and/or providing an ‘appliance.’ As someone who has worked in the retail industry for a long time, I think it's more a case of [both] figuring out how to differentiate so you don't dilute the ‘brand’ and [also] deciding if you want to be in the low-margin ‘price-conscious’ market segment at all," says Teradata admin Pluebell. "My company hasn't been interested in appliances to date; too limited in function and not suited to our more centralized administration model— we're still working to eliminate departmental data marts."
Elsewhere, data warehousing consultant Madsen cites the esoteric expertise that contributes to the expense of the Teradata experience, which he says frequently requires "a whole new—and more expensive—staff."
For this reason, other users—such as Teradata DBA Madson—say the soon-to-be independent company should focus on making Teradata Warehouse and its universe of associated tools easier to use.
For example, Madson says, he’d like to see "a more user-focused, rather than [a] technical[ly]-focused organization." That’s what’s really been holding Teradata back, he argues: "Teradata is designed by engineers for engineers. If they are going to expand their market they might benefit by creating a wider base of non-technical users of their systems and applications. More organizations might adopt Teradata if there were a broader pool of talent available to work with Teradata." In this last regard, Madson says: "The day I see a ‘Teradata for Dummies’ book will be the day I know Teradata is trying to further their penetration into the market."
Pluebell, for his part, says he isn’t so much concerned about expanding market share or lower pricing as he is with innovation. "Of course I'd like lower prices—but frankly the up-front system pricing isn't that far out of line with [non-appliance] competition for a large data warehousing platform, and Teradata generally leads in price/performance. And I would not want to give up all the good stuff just to get lower prices," he indicates, arguing, too, that Teradata’s current licensing model—of per-node, system-based client licenses—"generally favors customers" and is superior to the scheme it replaced.
That said, Pluebell adds, Teradata has a lot of interesting stuff in its R&D pipeline; hopefully, he says, the unfettered company will be in a better position to capitalize on these innovations. "I'd like to see them stay independent and aggressively move the technology forward. Teradata Engineering has lots of really interesting ideas—most of which I can't discuss, unfortunately, because I am bound by nondisclosure—and I'd like to see some of those become reality while I'm still around to take advantage of them," he concludes.