Related Technologies

A great deal needs to be done within DBMSes to make complex-data management more efficient. An example is sophisticated optimization algorithms to take advantage of new access methods and parallel processing. Yet, DBMS engines aren't the only component. Ancillary technologies must also evolve to help us create, store, maintain, and access data that does not fit neatly into rows and columns.

The processing performance necessary for handling complex-data types will come from multiple sources -- newer and faster processor designs, faster mem ory, higher-density storage, large-bandwidth networks, and OS enhancements.

Chip densities will continue to double every 18 months or so for the foreseeable future, which will help with the raw processing power needed for future data-processing demands. Similarly, 64-bit systems, such as those available from Sun, will arrive from IBM, the Hewlett-Packard/Intel partnership, and others. OSes with very large memory capabilities will benefit compute-intensive applications using complex-data types.

Similarly, symmetric multiprocessing (SMP), massively parallel processing (MPP), clustering, and hybrid-processor architectures will continue to make their presence felt, especially in data warehousing. At the same time, we'll see megabits of memory per chip, with sizes quadrupling about every three years. Magnetic-storage areal density (i.e., the number of bits stored in a given area of disk surface) under the newer magneto-resistive head technology is growing at a rate of 60 percent per year. Network technologies such as ISDN, Asymmetric Digital Subscriber Line (ADSL), and cable modems are evolving to satisfy the Internet's voracious appetite for bandwid th.

Nevertheless, there are at least two obstacles to overcome. One is the need for these technologies to conform to standards for interoperability. The second is achieving an acceptable total cost of ownership for these technologies.