Why Informix? And what about the coverage of midrange systems?
A reader ponders IBM's reasoning behind buying Informix; another questions <i>ES'</i> coverage of midrange systems ...
Why would IBM buy Informix? Except for IBM's Universe and UniData multi-dimensional databases, DB2 and the original Informix databases are very similar products. However, Universe and UniData, the two databases that Informix acquired when it bought Ardent software in early 2000, are multi-dimensional relational databases that are very popular for use in data warehousing projects. It therefore seems logical that those multi-dimensional relational database features are the ones IBM intends to integrate in future versions of DB2.
My reply is lengthy because this is a complex question of interest to Informix customers, investors and information technologists alike. In order to answer it, you have to place yourself in the position of an IBM executive rather than a technologist, because the business drivers are more compelling than any technology considerations. I have no inside information, but if I were an IBM executive ...
IBM has a great business opportunity with the Informix acquisition. First, GartnerGroup database marketshare figures for UNIX platforms in 2000 show Oracle with a commanding lead at 66 percent, double Oracle's marketshare on all platforms combined. IBM is a distant second with 14 percent of the UNIX market, followed by Informix with a mere 7 percent. Since 1998, Oracle's piece of this pie has grown from 61 to 66 percent. With the Informix acquisition, IBM's UNIX database marketshare climbs from 7 to 21 percent. Impressive.
Perhaps more importantly, Informix did not excel at cross-selling high margin services that contribute significantly to the bottom line at both Oracle and IBM. With IBM's vast portfolio of services, it may be able to reap higher profits from high margin services delivered to Informix customers like Verizon, Sears and American Airlines. At the same time, these value-added services will give IBM the opportunity to strengthen existing customer relationships where the company has been unable to penetrate the UNIX database area. It can also cross-sell traditional IBM products and services through relationships with newly acquired Informix customers. So far, IBM is successfully transitioning these relationships.
Another acquisition perk is a strong engineering group. Informix Extended Parallel Server (XPS) is a perennial player in the TPC-H benchmarks for databases 300 GBs and up. For example, it currently owns the highest Query-per-Hour (QphH) rating on HP servers in both the 300 GB and 1 TB categories. With some acquisitions, the staff of the acquired company starts looking for the nearest exit, but many of the Informix staff seem excited about the acquisition. At the close of 2002, if IBM still has 70 percent or more of the non-administrative staff they gain from the acquisition, I would consider it a personnel retention success and a positive indication of ongoing value from the acquisition.
On another point in your question: Though U2 (UniData and Universe) might be described as multi-dimensional relational databases for marketing purposes, they are more accurately described as multi-value relational databases, which have a loyal niche following of VARS and ISVs building small to mid-sized applications. Typically referred to as the embedded database market, the players are better defined by VAR-friendly business practices than by technology.
Without Informix, IBM has an insignificant presence in this market. With Informix, Big Blue becomes the third largest embedded database technology provider behind Progress and Sybase. While Microsoft is working hard to expand from its small- to mid-sized business base into the large enterprises that have favored IBM for decades, IBM could leverage the U2 technology to make inroads into the Microsoft small business applications fortress. There's more to the Informix database portfolio than UNIX marketshare alone if IBM can leverage it.
IBM will work hard to migrate users of the core Informix RDBMS products to DB2 in years to come. Oracle marketers are already working on strategies to lure these Informix customers. Time will tell if IBM can execute well enough to take advantage of the sizeable opportunity that the Informix acquisition represents.
Contributing Editor, Enterprise Systems
Cover That Midrange
Since starting to receive Enterprise Systems, I've noticed that the Midrange Systems supplement is growing thinner with each passing issue. Rather than offering just one or two general articles, it would be nice if the supplement could return to the original thrust of Midrange Systems, which provided insights on the AS/400 [now the IBM iSeries]. Specifically:
- What is IBM's real strategy for the iSeries, and how are sales holding up?
- Will an enhanced zOS for the zSeries supplant OS/400, or an enhanced AIX do likewise?
- If IBM follows through and puts Linux on all its four major server platforms and encapsulates them with a common GUI, then what (other than running legacy applications) will distinguish the iSeries from all the rest?
In short, give us articles that tell us where the iSeries is likely to be headed, even if it's to oblivion!
Thomas O. Potts
These are all great article ideas for readers with an ongoing investment in IBM's AS/400 and iSeries platform. To answer your question, Enterprise Systems will continue to cover that series, but not as a special section. Rather, we'll address iSeries and AS/400 issues through case studies, third-party product reviews and news items. That shift reflects our readers' need for information on a variety of platforms and operating systems, along with the integration and business issues that accompany them.