The new IBM moniker for data warehousing is Business Intelligence. While Business Intelligence may sound like an oxymoron in a Dilbertesque corporate world, and IBM’s acronym for it – BI – may seem to be an allusion to sexual proclivities, neither turns out to be true (well, not necessarily anyway). IBM’s marketeers simply couldn’t think […]
The new IBM moniker for data warehousing is Business Intelligence. While Business Intelligence may sound like an oxymoron in a Dilbertesque corporate world, and IBM’s acronym for it – BI – may seem to be an allusion to sexual proclivities, neither turns out to be true (well, not necessarily anyway). IBM’s marketeers simply couldn’t think of a more exciting term for data warehousing. (We can’t think of anything else either except perhaps data whorehousing, which can only appropriately describe the big black book that got Heidi Fleiss and Hollywood in so much trouble.) Nonetheless, the IBMers had to come up with some scheme to generate lots of press during the dull days of the first quarter, and thus the Business Intelligence initiative was born. At first glance, this new marketing campaign seems to be nothing more than a thinly veiled attempt by IBM to get its customers fired up about data warehousing so they’ll buy hardware, software and services. While this is certainly true – IBM’s AS/400, RS/6000 and S/390 server lines all saw significant revenue declines in 1997 – the Business Intelligence initiative, all kidding finally aside, is a serious effort on the part of IBM to not only help itself, but also to help customers better sift through the mountains of data they have stored on their servers to improve their businesses. The cornerstone of the Business Intelligence initiative is an expanded suite of Intelligent Miner programs. Intelligent Miner, which debuted initially on the RS/6000 in July 1996 and started shipping on AS/400s and S/390s six months later, provided a number of useful data mining algorithms that, up until that point, companies using IBM’s flagship DB2 data base had to buy from third parties or, more likely, create themselves. The tool also included some features to create graphical representations of data patterns. In February, IBM revamped the AIX implementation of Intelligent Miner. The latest IM suite ships first on the RS/6000 because that platform is by far IBM’s most popular for data warehouses, mainly because RS/6000s have the power and capacity of mainframes for a price that is consistently 20 percent below that of equivalent AS/400s.
By Timothy Prickett Morgan
With Intelligent Miner for Data version 2.1, IBM is embedding statistical analysis and forecasting programs right into the IM suite rather than making customers buy a third party stat package. The new IM also includes a new Java-based GUI, features to better exploit parallel data base queries on RS/6000 SP parallel servers (query parallelization makes complex queries run much faster) and a set of application programming interfaces to link Intelligent Miner to third party applications such as the popular ERP suites from SAP, JD Edwards and others. IBM says that version 2.1 also runs faster than the prior program. Intelligent Miner for AIX V2.1 costs $75,000, with support for additional nodes in SP parallel servers costing from $12,500 to $25,000 per node – the more nodes, the lower the per node price. On a big SP, customers could easily spend $500,000 to $1m on IM V2.1 alone, plus twice that for the data warehouse platform (hardware plus systems software). IM for Data V2.1 will be available on April 24; pre-configured RS/6000s (presumably with discounts) will be available June 12. Beta versions of the program will be available for OS/390 and Sun Solaris in April and on AS/400 and NT servers in June. IBM isn’t saying when the non-AIX platforms will be ready, but it will likely take six months to shake the quirks out of these implementations. IBM also isn’t saying what it will charge on these platforms, but customers should expect prices to rise as they have on the RS/6000s. IM for Data V1 cost from $15,000 to $42,000 on AS/400s and from $14,360 to $366,250 on S/390 servers (with significant price breaks for parallel Sysplex setups). Up until now, the Intelligent Miner tools have been limited to performing data searches and looking for trends in relational data base files. With the debut of Intelligent Miner for Text, a
new program that IBM calls version 2.1 for marketing’s sake, companies can now sift through customer correspondence, online news services, email and web pages and also analyze it for trends. The software can also be used to create web search engines for internet and intranet sites. IM for Text replaces IBM’s MediaMiner software, which has been previewing on its web site and which has been summarily withdrawn from marketing. As with IM for Data, IM for Text is shipping first on AIX. It will be available on March 27, which is the same date that the program will ship for NT servers. It will go into beta test on OS/390 and Sun Solaris in the third quarter; IBM has not decided to put it on the AS/400 as yet, but will do so if it gets wind that its big AS/400 customers want it. IM for Text V2.1 costs $30,000 for uniprocessor RS/6000s, with support for additional processors costing $10,000 each. IBM also improved its Visual Warehouse data extraction tools, which are used to suck information out of production databases and put them into data warehouses. Visual Warehouse 3.1 can be used to extract data from popular database management systems, including IBM’s IMS and VSAM flat file databases, the various flavors of its DB2 relational database, as well as Oracle, Microsoft SQL Server, Sybase and Informix data bases. The target warehouse at the other side of Visual Warehouse can reside in any DB2 family database or outside the DB2 family in a database accessible through DataJoiner. Prices for the software range from $15,000 to $67,000.