Welcome Guest. | Log In| Register | Membership Benefits
InformationWeek Labs

September 14, 1998


Microsoft And OLAP: Sudden Impact

continued...page 2 of 4

The strategy isn't guaranteed to succeed. A free offering may not be as compelling as it is in, for example, the browser world. Returns are much higher for P&DSS tools because they can quickly add value by making scarce and expensive developers and/or experts that much more productive.

Rivals' Reactions
Microsoft's strategy has triggered strong competitive reactions. Sybase matches all the Microsoft products with its new Warehouse Studio, and goes one better with Adaptive Server IQ and the Power Dimensions (WhiteLight) OLAP server providing both a relational OLAP (ROLAP) and OLAP engine, along with a more complete set of supporting warehouse tools. Oracle recently followed with its Data Mart Suite for Sales and Marketing and Business Intelligence packages to highlight its preprogrammed OLAP applications advantage. Broadbase, IBM, Information Builders, MineShare, Sagent, SAS, and others are polishing up their suites, awaiting word of Microsoft's final price, features, and performance to target their own offerings. These vendors will emphasize such unique advantages as scalability, cross-platform operating system capability, data mining tools, and advanced Web/ messaging functions along with the single-vendor responsibility and bargain pricing of a suite.

Meanwhile, the price and feature skirmishing between OLAP and front-end vendors has already been brisk. You almost need a scorecard to keep track of all the OLAP players, given the recent consolidations, mergers, and moves in the industry:
 
  • Arbor Software, in a $1.3 billion deal, merged with Hyperion Software to create the largest OLAP vendor, combining Arbor's leading MDD technology with strong vertical markets and P&DSS applications from Hyperion.
     
  • Cognos contracted with Panorama, the original designer of MS/SSOS, to market Aristotle, a front-end tool for MS/SSOS and any other OLE/DB for OLAP provider, worldwide.
     
  • Information Advantage bought cross-platform query-tool vendor IQ Software.
     
  • Hummingbird bought query and OLAP front-end tool vendor Andyne Software.
     
  • Seagate released a free OLE/DB for OLAP front-end program similar to Cognos' Aristotle called Seagate Worksheet, as a lead-in to its Holos and Crystal Info products.
     
  • Sybase bought the Intellidex repository and released Warehouse Studio package, an exact match to the SQL Server 7 data warehousing and OLAP bundle of products.

    Every vendor seems to welcome Microsoft to the P&DSS market, saying that its entry will "validate the market" or "create a new surge in business." For a market growing at 30% already, welcoming a rival's giveaway pricing is a strange rite of validation. But pricing is an issue in the P&DSS space. It's not that prices are high, but that they vary widely, even within each category. And pricing schemes are quite complex--they depend on operating system, the horsepower of the server, and what data sources you use. Given that data sources constantly shift in P&DSS, many users have to do a monthly reconciliation just to figure out how much to pay their P&DSS vendors.

    Single-product vendors cannot by themselves create a suite or bundle, and are most affected by Microsoft's giveaway. They're compelled to look for partners. Others shore up capabilities outside the classic OLAP space, hoping either to add value with new services, like MicroStrategy has with Broadcaster, or to serve the large top-end P&DSS market niche, where scalability, availability, and performance are key needs. Other vendors are adding complete applications and thus distinctive value to their products, like Gentia's Balanced Scorecard. There is a fair amount of room to maneuver in if the vendor pursues a value-added strategy.

    Design And Build
    Data warehousing projects can be multimillion-dollar money pits with little practical return. The Conference Board/Price Waterhouse study of data warehousing found that only one in four projects met or exceeded users' expectations. Some consistent management mistakes that create the problem include:
     
  • Underestimating the size and scope of the data reconciliation and cleaning tasks;
     
  • Focusing too much on warehouse infrastructure and too little on analysis and planning processes;
     
  • Getting lost in the details of hundreds of databases and ever-changing demands;
     
  • Promising fixed costs against a backdrop of constantly shifting needs;
     
  • Having to cope with managing new operational systems, which pass outside traditional organizational boundaries.

    Perhaps the most persistent problem is that data warehouse teams never quite get beyond designing and building. Consistently, 10% to 25% of their time is devoted to new projects, problems, or requirements. Good "design and build" tools for data warehousing are critical to success.

    Data modeling tools are vital to data warehouse designing and building for two reasons. First, their reverse-engineering capabilities let designers view all the tables and fields of most databases in detail but also in Entity Relationship diagrams. You can use this information to more quickly design new data warehouses or data marts, and to quickly spot and deal with data-field mismatches--for example, customer name fields from two tables having different lengths or layouts--or coding inconsistencies.

    Second, you can use them for "forward engineering"--after you design a new data warehouse from old, new, and/or amended data fields, you can generate it automatically and add in trial data for immediate testing. Some data modeling tools provide estimates of the size of the new data warehouse based on the new data mart's table and field sizes plus the constituent tables' number of records. Others are able to synchronize a design by comparing all the tables and fields in an existing database with the planned new design and highlighting all the changes and differences (expected and unexpected). Finally, some of the best data modeling tools automatically create data migration scripts one can use to transfer data from constituent tables over to the new data warehouse or data mart.

    Microsoft's Visual Data Modeler's features are serviceable, but by no means top-of-the-line. Although it can inspect many databases, you can use Data Modeler only to reverse-engineer SQL Server and Oracle tables. On the forward-engineering side, it works only with SQL Server databases--without the extra features such as synchronization or size estimation. In sum, Visual Data Modeler is targeted specifically as a SQL Server 7 tool, and does very good job at that. It won't replace the functions of other tools like Anubis Constructa, Powersoft/Sybase Warehouse Architect or Platinum/LogicWorks ERWin.

    continued...page 3, 4
    return to page 1



    Back to Labs

    Send Us Your Feedback

    Top of the Page
    CAREER CENTER
    Ready to take that job and shove it?



    TechCareers

    SEARCH
    Function:

    Keyword(s):

    State:
  • Post Your Resume
  • Employers Area
  • News & Features
  • Blogs & Forums
  • Career Resources

    Browse By:
    State | City
  • SPONSOR
    RECENT JOB POSTINGS
    CAREER NEWS
    Go beyond Google and get vertical. These specialized search sites will help you find the business information you need -- fast.

    Ari Balogh was named to the post of chief technology officer as the companys for a "realignment" of employees.



    Specialty Resources

    Featured Microsite