Welcome Guest. | Log In| Register | Membership Benefits

News In Review

June 14, 1999

Print this story
Print this story
Data Capture Grows Wider

continued...page 2 of 3

Illustration by James Yang
Related links:
  • Excelon Stores And Delivers XML Data

  • Tapping The Pipeline

  • Profitable Customers
  • And from our sister publications:
  • EETimes High-speed net, computers open data frontier
  • Oracle this month will begin shipping Oracle8i Lite, the latest release of its lightweight database. Oracle8i Lite can be as small as 50 Kbytes, says Denise Lahey, VP of Oracle's mobile- and embedded-products division. By comparison, Oracle's Personal Oracle7 database was 8 Mbytes--too big for most portable computers, let alone PDAs or embedded devices.

    Stay In Control
    Beth Israel Deaconess Medical Center in Boston is using computerized devices to manage drug distribution to patients. The cabinet-like MedStation Rx systems, each with a built-in PC board, Sybase database, and flat-panel display, are located around the hospital and are stocked with medication. "The chief benefit has been better medication inventory control," says Christopher Schifferdecker, a pharmacy manager at the medical center. "It was a very labor-intensive process."

    The devices signal to a central pharmacy when they need to be refilled and generate patient-billing information. The system is capable of automatically reordering drugs from distributors, though Beth Israel Deaconess is not yet doing that.

    It's projects like these that have IBM bulking up the resources it devotes to pervasive computing. Last year, IBM assigned about 1,000 employees to work on pervasive computing, and in May, it added another 550, according to Mark Bregman, general manager of pervasive computing at IBM.

    IBM sees three markets developing in pervasive computing: hardware and software components for device manufacturers; consulting and development work to support business customers; and systems for service providers building security, billing, and management infrastructures. In the latter category, IBM and France Telecom will launch a pilot in France in a few weeks that involves a "service-delivery platform" for voice-and-data screen phones. Bregman says the platform will be available next year.

    Last month, IBM unveiled a beta version of VisualAge for Embedded Systems, a development tool for building embedded applications in Java that are smaller than 1 Mbyte. IBM also unveiled a version of its relational database for Windows CE and Palm OS devices, called DB2 Everywhere. Both products are designed to connect portable and embedded devices with back-end systems.

    Millions Of Devices
    From a data-analysis standpoint, the real potential in pervasive computing may lie in a company's ability to centrally manage and mine the data generated by mobile and embedded devices. Can data warehousing and data mining--technologies primarily used to manage data generated by back-end business systems--be brought to bear on thousands, or even millions, of smaller devices?

    "This gets to the foundation of what data warehousing and business intelligence are all about," says Michael Burwen, president of the Palo Alto Management Group, a market-research firm. Retail chains, insurance firms, and shipping companies generally represent the leading edge in leveraging data from mobile and embedded systems for data-analysis tasks, Burwen adds.

    Federal Express collects customer-transaction data from its supertracker and airbill systems, as well as from FedEx's Web site, and loads it into a 2-terabyte data warehouse, which feeds summary data into departmental data marts for analysis. A 600-Gbyte marketing data mart is used to determine which services FedEx offers to different customers.

    "We've gotten out of the realm of mass marketing," says Tom Wicinski, FedEx's marketing analysis managing director. "This really allows us to target our customers." FedEx uses an Informix database to power its data warehouse and Oracle databases for its marketing, sales, and other departmental data marts. "Having all this data has helped us build the right predictive models," Wicinski says.

    Cummins Engine Co. began building electronic components into its engines years ago for gathering information about performance, pollution control, and other diagnostic data. Only recently, however, has the Columbus, Ind., manufacturer begun experimenting with collecting and mining that valuable data. For example, Cummins found that one customer's trucks were idling three times as long as trucks at other companies, leading to more wear and tear on the engine, higher fuel consumption, and more pollution.

    continued...page 3
    return to page 1


    Illustration by James Yang


    Back to This Week's Issue

    Send Us Your Feedback

    Top of the Page