email article printer friendly more resources

Database Management System Market Dynamics
Why DBMSs have a resurgent role in the application platform landscape.
by Peter O'Kelly

October 15, 2004

Database management systems (DBMSs) have a fundamental role in application platforms, but there is currently a lot of market confusion about how, when, where, and why DBMSs should be used. This column provides an overview of DBMS trends and the reasons why DBMSs have a resurgent and expanding role in the broader application platform landscape. The next Trends & Analysis column will assess how Microsoft's SQL Server 2005 product family fits into both the emerging DBMS market landscape and Microsoft's overall Windows Server System strategy.

A Brief History of DBMSs
To establish context, it's useful to briefly review the fundamental DBMS value proposition. DBMSs are used to securely and robustly manage databases; databases are sets of data that capture descriptions of real-world things such as customers and product inventory. DBMSs embody sophisticated technology to efficiently and concurrently make data available to applications and users without compromising database integrity.

The DBMS market has evolved through several generations, starting with hierarchical DBMS products such as IBM's IMS, which was introduced during the late 1960s. Network DBMS products (also known as CODASYL) came along next, with Cullinet's IDMS (today a Computer Associates "legacy" DBMS) serving as a leading example. Relational DBMSs entered the market approximately 25 years ago and, after a phase of "database wars" between relational and network DBMS products, have dominated the DBMS market for most of the past 20 years. IMS and IDMS are still used for legacy applications, but today the vast majority of database developers work with relational DBMSs (and hereafter "DBMS" refers to relational DBMS unless otherwise noted).

DBMS evolution led to higher levels of abstraction for database developers and users. While hierarchical and network DBMSs presented models that mixed logical and physical (implementation-oriented) considerations, for example, developers and users working with relational DBMSs focus primarily on logical abstractions that more closely mirror the real world things and events described by the database; they aren't conceptually burdened with pointer chains, buffer pools, and other low-level details. As a result, DBMSs are conducive to improved developer productivity as well as increased overall system security and robustness.

Perennial DBMS Challenges
DBMSs offer more secure, robust, and productive options for data management, so why isn't all data stored in DBMSs today? It's estimated that 70 percent or more of most organizational data isn't currently stored in DBMSs, but is instead scattered across file systems, e-mail messages, and assorted specialized content/document management systems.

Several longstanding challenges have prevented broader DBMS applicability and, as we'll see momentarily, the advent of Web-centric applications was in some respects a further setback for DBMSs. One major DBMS challenge, historically, had to do with cost and complexity. DBMSs, especially for high-end systems, were expensive to license and maintain. They required DBMS-trained developers and administrators, and also required extensive fine-tuning, which in turn resulted in protracted application development and test cycles. DBMSs also have more demanding hardware requirements than simpler, "good enough," file system-based alternatives, and that was a key consideration during the 1980s and 1990s, when hardware was still expensive relative to today's market.