Printer Friendly

Trends in database management systems technology.

In an age when enterprises (and individuals) are handling more and more information, managing this data efficiently and securely has become a challenge. And with every challenge, innovators can step up to the plate with technology that makes the process of storing, modifying, and extracting data almost painless, whether the database is a computerized library system, automated teller machine network, or a remote inventory system.

As database management system (DBMS) vendors continue to enhance their products, it's becoming more critical to provide the right information at the right time to the right people, according to Donald Feinberg, vice president and analyst for Stamford, Conn.-based Gartner, Inc.

Just take a look at the increased investments to see the growing importance of database management systems. The DBMS market grew 10.1 percent, from $8.7 billion in 2003 to $9.6 billion in 2004; and the relational database management system (RDBMS) market grew 10.2 percent from $7.1 billion to $7.8 billion in new license sales over the same time frame.

"This is greater growth than the overall IT budget increase worldwide for the same time frame," Feinberg said. "Companies are still buying new and additional DBMS engine licenses, and we believe this trend will continue during our visionary horizon of [5] years."

The expanded use of data warehouses across enterprises, according to Feinberg, is driving the growth. Atone time, only information managers and technicians used data warehouses. But now, marketing, accounting, and other departments across the enterprise are accessing data.

By 2007, Feinberg expects half of all data warehouses to use closed loop systems that are continuously updated: As business users access data from the warehouse and for their own needs, any steps they take (i.e., implementing a new marketing program) will be fed back immediately into the data warehouse source systems.

Although companies using data warehouses have already grown, smaller markets are expanding, said Feinberg. "By 2007, due to increased speed of hardware and software, small and midsize [business] implementations of OLTP [Online Transaction Processing] and data warehouses will begin to merge to a single database with shared schema. For large enterprises, the use of a single database for both OLTP and data warehouses will not be possible until after 2010."

As users become more comfortable with the systems, the complexity of the queries puts more of a strain on the system's resources than the actual number of queries, according to Feinberg. This means that companies often underestimate the processing power they will need for their data warehouse management systems.

As query complexity increases, the data warehouse structure must simplify the process and may require some data warehouse appliances for peak performance, according to Feinberg.

Using XML will be one way companies can simplify the data warehouse structure, according to Feinberg. He predicts that as major DBMS vendors add native XML and XQuery leading to XML data in the DBMS, XML will become the majority portion of the data in the warehouse by 2008. Feinberg recommends that companies start to move their XML data to the DBMS engine for reliability, availability, and security.

More Data in More Formats

Feinberg cautions that data warehouses are being asked to store increasing amounts of disparate data in an increasing variety of formats. This includes data such as photos, music, and scanned records and faxes, as well as other data with meaning to specific types of applications (i.e., medical records, security and biometric data, and geodetic data).

"As we move to store all company data in a single data store, standard search engines must be able to use this data in searching," said Feinberg. The database management system models of today will continue to be the standards for growth in the next few years. Through 2010, the predominant model for new, emerging information management technologies will be via extensions to and evolutions of the RDBMS model, according to Feinberg. ("Relational" models refer to the way the DBMS organizes information internally.)

Open Source Adoption Growing

One of the major evolutions that Feinberg sees is a wider adoption of open source systems (OSS), which are just starting to gain traction. Few organizations use OSS to monitor, analyze, and manage a network system in their data warehouse systems today. This is par tially because most DBMS engines don't include OSS support. Few tools actually exist to manage large database implementations of OSS.

Even with momentum growing for OSS, however, Windows-based systems aren't going away any time soon, said Feinberg. He also cautioned that DBMS engines based on open source code are still in their infancy.

So companies that may use them should keep close control of these engines to ensure they will perform as expected.

One of the draws for all open source-based applications--not just those involving data warehouse and database engines--is free support from the open source community. Feinberg warned that the free help may do little, if anything, to solve any implementation or operational issues.

"Always purchase support and maintenance from the vendor," said Feinberg. "Also be prepared to use internal resources for support.

"However, the next few years as the demand increases, we expect to see the software vendors increase support for these engines," Feinberg said. Yet these implementations will tend to be with smaller databases, typically under five terabytes. So Feinberg recommends limited use of DMBS engines based on OSS for simple tasks, such as supporting Web servers. "OSS DMBS is not yet ready for mission-critical applications," according to Feinberg.

Vendor Outlook

Though the market is still dominated by traditional players including IBM, Oracle, Teradata, and Microsoft, the market demand is driving many new vendors in the market--CopperEye, DATAllegro, Metapa, Netezza, Sleepycat Software, and more, according to Feinberg, who expects this trend to continue. A few other companies, including Informix, Ingres, Red Brick, SAP DB, Supra, and TimesTen have left the market.

Feinberg cautioned that though some of the new market entrants have strong business intelligence backgrounds, which can be important as to support more complex querying needs, they may not provide all of the DBMS technologies needed to support a particular enterprise's short and long-term data warehousing needs.

Some of the traditional players are focusing on their core capabilities to strengthen their market positions, Feinberg added. For example, Oracle has typically been strong in DBMS, while Microsoft has been dominant in operating systems.

"By 2008, Microsoft will be a strong contender in data marts on the Windows platform, but [it will] still be restricted by a lack of support for other operating systems," said Feinberg. Meanwhile, Oracle will focus on strengthening its Linux operating system capability to offer an open source alternative to the proprietary Microsoft platform.

"The database wars of old will be waged on new ground. Instead of functionality wars fought at tactical levels (for example, who has a better SQL [structured query language, the favored language for DBMS systems] or better replication), they'll now be fought at higher and more strategic levels," said Feinberg. "Through 2007, there will be significant differences among the capabilities of various DBMS products to effectively support large data warehouse implementations."

Through 2008, Feinberg expects NCR's Teradata to maintain its high-end data warehouse capability lead. But that dominance will be more difficult to defend. Teradata also lacks the marketing prowess to change the market rules and promote the company's midrange capabilities, he said. Microsoft serves the small and midsize market with SQL server, but this application doesn't perform very well in larger installations, according to Feinberg.

We'll keep watching for the enterprise that steps up to the plate with the next-generation innovation for DBMS.

Phillip Britt, president and CEO of S&P Enterprises, Inc., is a business writer who covers key topics in the information industry. His e-mail is spenterprises@wowway.com. Send your comments about this article to itletters@infotoday.com.

Tips for Users

* Examine data warehouse implementations that are closed loop--i.e., those that feed data back to source systems to tune and refine operational processes.

* Develop skill sets for Linux as a production platform, especially for DBMS.

* Use commodity hardware and build-out architectures to reduce costs.

* Experiment with open source DBMS engines for data warehousing.

* Examine your vendor's overall architecture and how it relates to your information architecture.

Donald Feinberg, vice president and analyst for Gartner, Inc., offers these suggestions for data users:
COPYRIGHT 2006 Information Today, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Britt, Phillip
Publication:Information Today
Geographic Code:1USA
Date:May 1, 2006
Words:1386
Previous Article:Exalead offers unified search experience.
Next Article:Push to register for .eu domain name.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters