By Mark Madsen, Intelligent Enterprise
I was sorry to see last week's Teradata Partners user conference come to a close. The event was half Teradata technical sessions and half business sessions, where I spent more of my time. The business sessions were roughly equal parts enterprise BI platform and methods, data management topics, customer and business process BI, and active data warehousing. There were other topics mixed in as well, including a number of sessions that highlight the deepening company relationship with SAS.
While not a stated theme, I noticed an increase in Web data as the subject of analysis, or married to internal transactional and customer data. There was even a session on analyzing social networks based on cellular call data — not exactly Web, yet a topic most commonly associated with Web businesses. There were presentations by eBay, PayPal, and Netflix as well as non-Web companies and government departments.
Rich internet applications (RIA) and BI: the user interface is no longer an important differentiator between BI tools and general programming toolkits. In fact, BI tool UIs are generally weak on interactivity compared to custom-built counterparts. I saw several examples of companies using Flash or Ajax with user interfaces that blow away what the big BI vendors are delivering. The real strength of BI tools is in their metadata and query generation layers. It's hard to duplicate that in a custom-built Web interface for BI users, and I would say it's the one area holding many people back from doing more custom BI work.
Open source increased in visibility over last year, as it has across the industry. Novell, Talend and Sun had booths here. Talend released a very low-key partnership announcement with Teradata at the event. Vincent Pineau and Yves de Montcheuil discussed progress with their Teradata-specific SQL generation and ELT options.
Several companies discussed using open source tools to build custom interfaces; one of the customers was using Jaspersoft as a key element in their reporting systems. In one of the sessions someone from Teradata commented on their shifting from proprietary Unix to Linux and stated that Linux is equivalent in reliability and security to proprietary Unix, putting to rest any lingering doubts about Linux for large database workloads.
Last month, rumors were floating around (thanks to someone from JMP Securities) that SAP might be considering a Teradata acquisition. The people I talked to at Teradata said the rumors were not true. I believe them more than the rumor because of current fit and market conditions, but then I didn't believe Microsoft would buy DATAllegro over some of the stronger appliance offerings out there, so take that for what it's worth. On that front, I heard a (still unsubstantiated) rumor that Microsoft reduced its price for DATAllegro after digging further into the technology. I'm not sure I believe this particular rumor, but I heard it from more than one person so I'm trying to find out more.
For me the biggest surprise was how far Teradata has shifted from their dogmatic view that an enterprise data warehouse (EDW) is the only solution to BI problems and that you therefore only need one type of platform to meet these needs.
Two things have been shifting this view. One is the reality that we can't always store 100% of our enterprise data. There are external data needs, incompatible perspectives of details and special cases that come up constantly and require more flexibility than a single central EDW model can provide.
I believe the bigger factor is the market realities for business intelligence. There is the simple fact that many businesses build and deploy BI at the division or department level. These are more limited in scope and budget. An EDW is something that can only evolve in these businesses, if it happens at all. By ceding this market to database and appliance vendors, Teradata was limiting its potential customer base.
At the same time, there are many companies attempting to eat into Teradata's business from below, a bad position for any vendor to be in when the product segment is becoming a commodity. To stay competitive when the market is paying so much attention to appliances meant that Teradata had to introduce a broader product line. They needed to give people without an EDW or huge data volumes the products that would suit their more limited needs and do this for a more competitive price.
The new product lines mentioned in an earlier post and moderating the EDW message to one that is more pragmatic were important shifts in company strategy. This change has been a challenge. A lot of people have been with Teradata for years and have lived the EDW dogma. Changing these beliefs and practices doesn't happen overnight, and that slowed the product introductions and sales of the new product lines. Hopefully enough people within the company will recognize that this is driven by market realities, not religious debates, and pragmatism will prevail.
Source: Intelligent Enterprise
About the Author
Mark Madsen is a Contributing Editor with Intelligent Enterprise and is president of Third Nature, a consulting and research firm focused on business intelligence, data integration and data management. He is a principal author of Clickstream Data Warehousing and speaks about data warehousing and emerging technology. Write him at firstname.lastname@example.org.