A Brief History of DB2 for Linux, Unix, and Windows by Chris Fierros

Although roughly ninety percent of DB2 Universal Database (UDB) on distributed platforms such as Windows , Linux, and UNIX is common code, IBM makes great efforts to exploit the features and technologies of each operating system it supports. In fact, DB2 UDB v8 for Windows leverages more features and technologies unique to the Microsoft Windows operating system than any other supported platform.

It’s probably unlikely that you have been working with DB2 Universal Database (UDB) on distributed systems since the original Database Manager version 1.0 was announced in late 1987, which was almost a decade before it was announced for on the Windows NT 4.0 platform and long before it was available on AIX which was the first UNIX based platform announced by IBM. For the reader that has not spent the last decade working with the DB2 Universal Database on Intel platforms, here is a brief history of the long road that has lead us to the most significant point release to date, DB2 UDB v8.2 for Windows.

DB2 Universal Database (UDB) v8.2 is IBM’s latest release of the relational database management solution which has achieved worldwide market share leadership on a wide variety of platforms including the Windows operating system. DB2 UDB provides customers of all sizes with remarkable value by implementing innovative self-management and automation technology combined with superior price performance and scalability.

DB2 Universal Database has been available on Intel platforms since 1987. The product was originally developed for the 16-bit OS/2 and included in the packaging of OS/2 Extended Edition v1.0. The product was then called Database Manager (DBM) and the name as well as the acronym can still be found in many aspects of the product today, for example multiple DB2 UDB instances today are equivalent to what was once a single Database Manager And today we still configure these instances (database managers) using commands like GET DBM CFG and UPDATE DBM CFG. The next release of Database Manager for OS/2 EE v1.1 shipped with an API for compiler vendors to develop there own pre-compilers. It also delivered Structured Query Language (SQL) operators enhancements including: UNION, INTERSECT, AND EXECT as well as a CURSOR HOLD OPTION which would give application developers the ability to hold a cursor position across transactions (or unit of work) commit points.

In late 1991 the product was ported to 32-bit for OS/2 v2.0 and included in the packaging of OS/2 Extended Services v1.0. In early 1993 the product was renamed to DB2 for OS/2 (DB2/2) version 1.0 and development was moved to Toronto, Canada. A couple of years later, DB2 for OS/2 was ported to AIX and the product was renamed DB2 Common Server to indicate a common code base among the distributed platforms. It was with DB2 Common Server that porting to other Intel and UNIX based operating systems began as announcements were made for beta versions to support HP-UX, SUN Solaris, and Microsoft Windows NT. Online backups were first introduced with DB2/2 v1.2 which would allow database administrators to perform database backups while applications remained connected to the database.

DB2 for Windows NT Version 2.1 was announced towards the end of 1995. The following is a quote from the original DB2 for NT announcement:

“If you need a reliable and powerful database on Windows NT in a stand-alone, departmental, or enterprise environment, IBM's premier family of relational database products, DB2, is now available.”

In 1997 IBM announced the next generation of relational database technology and the name DB2 Common Server was changed DB2 Universal Database. The first DB2 Universal Database, version 5.0, made the Extended Enterprise Edition (EEE) available on the Windows NT operating system as well as providing support for the native Named Pipe communications protocol. DB2 UDB v5.2 provided support for the Virtual Interface Architecture as a supported Fast Communications Manager protocol to enable high-speed inter node communications between physically partitioned database nodes.

In September of 2002 IBM announced DB2 UDB v8.1 marking the next stage in the evolution of the relational database management system with a focus on Self Management and Resource Tuning (SMART) features which included the ability to define AUTOMATIC instance and database configuration parameters values. Version 8.1 also shipped with several high availability enhancements including online loads, reorgs, buffer pools, and configuration parameters.

DB2 UDB v8.2, code named "Stinger", includes a number of new features including enhanced reliability, manageability, integration and scalability features. This point release represents close to a decade of relational database development effort by IBM on the Windows operating system platform and is perhaps the most significant. Probably the most significant new feature in DB2 UDB v8.2 is High Availability Disaster Recovery. HADR is a DB2 feature that integrates both high availability and disaster recovery by replicating data, via log buffer shipping, from a primary database to a standby database. HADR provides data protection for both partial system failures due to hardware or software failures and complete site failures due to natural disasters. Read more about High Availability Disaster Recovery in this info brief.

In June of 2006 IBM announced DB2 v9.1 IBM announced DB2 v9.1, code named “Viper”, as the "the next -generation hybrid data server with optimized management of both XML and relational data". DB2 v9.1 is the first version of DB2 that did not include the name Universal Database (UDB) in over a decade. This version included some significant deployment enhancements over DB2 UDB v8.2 including; a new option to install DB2 on Windows without administrator privileges, but more significantly was the ability to install multiple copies of DB2 on a single server which allows v8.2 and v9.1, in addition to future version, to coexist on a single server. This feature reduces the cost and increase the ease of migrating to DB2 v9.1. A very significant enhancement to security in came the form of label based access which allows more granular control of access to both rows and columns which also includes a new database level privilege of Security Administrator. Finally, DB2 v9.1 also included a feature called table partitioning which allows for easier management of very large tables by allowing them to be partitioned into smaller more manageable tables that can quickly be rolled-in and rolled-out of the partitioned table.

In October of 2007 IBM announced DB2 v9.5 providing a number of enhancements to features already in DB2 v9.1. The High Availability Disaster Recovery (HADR) feature originally introduced in DB2 UDB v8.2 has been integrated with Tivoli System Automation to provide automated fail over of HADR databases. Significant enhancements in database security auditing were made to the DB2 Audit Facility which allows for more granular security auditing and easier historical data archival.

In May of 2009 IBM announced DB2 v9.7 providing even more enhancements to features in already in DB2 v9 .5.The High Availability Disaster Recovery (HADR) feature now supports read only work loads on the standby HADR database.