IBM Db2

IBM Db2
Company type	Product line
Industry	Database management software
Predecessor	IBM SQL/DS
Founded	1983
Founder	IBM
Headquarters	Armonk, New York, United States
Area served	Worldwide
Products	Db2 database servers and data management solutions
Owner	IBM
Parent	IBM
Website	www.ibm.com/analytics/db2

Db2 represents a suite of data management solutions, including various database server products, all developed by IBM. Initially designed to support the relational model, its capabilities have since grown to incorporate object–relational functionalities and accommodate non-relational data types such as JSON and XML. The product name was officially written as DB2 until 2017, when it transitioned to its current styling. Early mentions occasionally misrepresented it as DB/2, erroneously linking it to the OS/2 operating system.

History

The inception of DB2 traces back to the early 1970s following IBM researcher Edgar F. Codd’s introduction of relational database theory, which laid the groundwork for data manipulation models in 1970. In 1974, IBM’s San Jose Research Center developed System R, a Database Management System that implemented Codd’s theoretical framework and gave rise to Structured Query Language (SQL). At first, IBM did not fully adopt Codd’s concepts, leading to the creation of Structured English QUEry Language (SEQUEL), an early iteration of SQL.

To enhance its graphical user interface and SQL capabilities, IBM acquired Metaphor Computer Systems, whose technology had been in use since the mid-1980s. Simultaneously, IBM introduced Query by Example (QBE), recognized as the first graphical query language.

IBM’s initial commercial relational database product, SQL/DS, was launched in 1981 targeting DOS/VSE and VM/CMS platforms. Earlier, in 1976, Query by Example was released for VM, offering a table-oriented front end that generated commands in a linear syntax for relational databases. Subsequently, DB2 integrated QMF, providing full SQL functionality with an interface reminiscent of QBE. The architectural design of DB2 drew heavily from IBM’s hierarchical IMS and its associated database language DL/I.

The DB2 brand was introduced in 1983 with its release on IBM’s MVS mainframe system, marking a transition from hierarchical databases such as IMS to relational database technology. IBM’s promotion of SQL through DB2 played a significant role in its widespread acceptance across the industry, overshadowing alternatives like the QUEL language used by Ingres. DB2 became generally accessible in 1985 and by 1989 generated revenues nearing $1 billion, rivaling those of IMS.

Originally limited to IBM mainframe environments, DB2 broadened its reach during the 1990s to include platforms like OS/2, UNIX, Windows, and Linux (including Linux on IBM Z), as well as personal digital assistants (PDAs). Additional versions were developed for z/VSE and z/VM. Early iterations of DB2 LUW (Linux, Unix, Windows) were derived from OS/2’s Database Manager Extended Edition codebase.

IBM expanded Database Manager’s functionality by introducing distributed database features through Distributed Relational Database Architecture, which allowed shared remote access over LANs using protocols originating from the Distributed Data Management Architecture (DDM).

In later stages, IBM redeveloped the software and released updated editions named DB2/2 and DB2/6000. This naming pattern extended similarly to versions such as DB2/400 (for AS/400), DB2/VSE (for DOS/VSE), and DB2/VM (for VM OS). Due to legal considerations, product names were eventually revised to follow a "product FOR platform" format (e.g., DB2 for OS/390). The subsequent generation was marketed under the name DB2 Universal Database (DB2 UDB).

Db2 for LUW

During the mid-1990s, IBM introduced a clustered version of DB2 known as DB2 Parallel Edition, initially available on AIX systems. This variant facilitated scalability through a shared-nothing architecture, distributing a large database across multiple DB2 servers connected via a high-speed interconnect. Subsequently, IBM extended support to Linux, UNIX, and Windows (LUW) platforms, renaming it DB2 Extended Enterprise Edition (EEE). Today, this functionality is referred to as the Database Partitioning Feature (DPF) and is included with IBM's premier DB2 Enterprise product.

Following Informix Corporation’s acquisition of Illustra and its transformation into an object-SQL DBMS with Universal Server, both IBM and Oracle Corporation adapted their database engines to incorporate object–relational capabilities. IBM’s 2001 purchase of Informix Software enabled the integration of Informix technologies into DB2, effectively positioning it as an object–SQL DBMS.

In 2006, IBM unveiled "Viper," the codename for DB2 9 targeting distributed platforms and z/OS—the latter announced in early 2007. This version pioneered native XML storage within a relational database, introduced enhancements for OLTP on distributed systems, improved business intelligence and data warehousing on z/OS, expanded self-managing features, increased 64-bit platform support, and boosted stored procedure efficiency.

October 2007 marked the release of "Viper 2" or DB2 9.5 for distributed environments, focusing on Simplified Management, Business Critical Reliability, and Agile XML development capabilities.

In June 2009, IBM launched "Cobra," known as DB2 9.7 for LUW. This update incorporated data compression for indexes, temporary tables, and large objects; enhanced native XML support within hash and range partitioning; and introduced multi-dimensional clustering advantageous for data warehousing scenarios. Additionally, DB2 9.7 increased compatibility with Oracle Database by supporting common SQL syntax, PL/SQL, scripting languages, and Oracle-specific data types. The concurrency model was also refined to mirror behaviors found in Oracle and Microsoft SQL Server.

October 2009 saw the introduction of DB2 pureScale—a clustered database solution designed for non-mainframe OLTP workloads based on the mainframe Parallel Sysplex architecture. pureScale delivers fault tolerance, shared-disk storage capabilities, scalability up to 128 nodes, continuous availability, and automatic load balancing.

Also in 2009, DB2 was integrated as an engine within MySQL, allowing users on IBM i and other platforms to access DB2 files through MySQL interfaces. On IBM i (formerly OS/400), DB2 is deeply embedded within the operating system and supports features such as journaling and triggers.

Early 2012 brought DB2 10.1 (codenamed Galileo) for Linux, UNIX, and Windows platforms. This release added advanced data management features including row- and column-level access control for granular security as well as multi-temperature data management to optimize storage costs by categorizing data based on access frequency. It also introduced "adaptive compression" to improve table data compression efficiency.

IBM followed with the release of DB2 10.5 ("Kepler") in June 2013.

On April 12, 2016, IBM announced the availability of DB2 LUW 11.1, which became generally available in June 2016.

By mid-2017, IBM consolidated its DB2 and dashDB products under a single "Db2" brand identity.

On June 27, 2019, IBM launched Db2 11.5—the AI Database—integrating artificial intelligence features aimed at enhancing query performance and supporting AI-driven application development workflows.

Db2 for z/OS - progression

The 1995 release of Version 4 introduced data sharing, enabling multiple DB2 engines to concurrently access identical data, thereby enhancing both performance and system availability.

Version 5, launched in 1997, brought the capability for online reorganization of tablespaces.

In 1999, Version 6 added support for object-relational data types exceeding 32K in size—such as images and videos—and incorporated trigger functionality.

Version 7, released in 2001, featured dynamic data set allocation, utilities applicable to tablespace lists, real-time statistical monitoring, scrollable cursors, and the initial introduction of Unicode support.

With the 2004 release of Version 8, DB2 gained 64-bit architecture compatibility, new index varieties including the notable DPSI index, recursive SQL capabilities, and transitioned its internal catalog to Unicode format.

Version 9 arrived in 2007 with Trusted Context security enhancements and native XML handling.

In 2010, Version 10 introduced Temporal Tables for tracking row history, strengthened security through separation of system and security administrator roles, and implemented row and column access control (RCAC).

Version 11, released in 2013, incorporated JSON data support.

The 2016 Version 12 update added RESTful service interfaces and AI-driven optimizations for data access paths aimed at boosting performance efficiency.

IBM unveiled Db2 13 for z/OS on May 31, 2022, delivering advanced AI capabilities alongside improvements designed to solidify its position within enterprise hybrid cloud and digital computing landscapes.

Db2 Warehouse

IBM first conceptualized "data warehousing" in 1988 to overcome the shortcomings of traditional transactional databases when extracting business insights. Db2 Warehouse was engineered to rapidly consolidate and analyze data from numerous transactional sources at scale.

As data volumes expanded exponentially—from gigabytes to petabytes—Db2 Warehouse evolved beyond conventional star and snowflake schema designs to accommodate a broad spectrum of data formats and analytical frameworks including:

Both relational and non-relational datasets
XML alongside geospatial information
Integration with RStudio and Apache Spark platforms
An embedded Spark analytics engine
Massively parallel processing architectures
In-memory analytics processing
Predictive modeling techniques

Operating within Docker containers, Db2 Warehouse supports deployment across on-premises infrastructure, private clouds, and various public cloud providers. It is offered as either software-only or appliance-based solutions compatible with Intel x86, Linux systems, and mainframe environments. Utilizing IBM’s Common SQL engine enables efficient querying across diverse data sources—such as Oracle, Microsoft SQL Server, Teradata, open-source databases, and Netezza—allowing users to execute unified queries that rapidly return integrated results.

Db2 on Cloud / Db2 Hosted

Db2 on Cloud (previously known as "dashDB for Transactions") is a fully managed cloud-based SQL database service providing a 99.99% uptime SLA with independent scalability of compute resources and storage capacity. It supports seamless rolling security updates and can be deployed on both IBM Cloud and Amazon Web Services (AWS).

Its principal business advantages encompass:

Elastic scaling: Compute resources can be adjusted independently to accommodate peak workloads while storage scales separately based on demand.
Backup and recovery: Maintains backups for up to 14 days with point-in-time restore capability and offers one-click failover to offsite disaster recovery sites.
Encryption compliance: Adheres to data protection regulations featuring encryption at rest, SSL-secured connections, continuous security updates, and daily backup routines.
High availability: Utilizes Db2’s HADR technology to ensure 99.99% uptime while allowing maintenance and scaling without service interruption.
Data federation: Supports integrated querying across Db2 instances whether on-premises or in cloud environments.
Private networking: Enables deployment within isolated network segments accessible via secure VPN connections.

Db2 Hosted (formerly "DB2 on Cloud") represents an unmanaged hosted solution providing direct server installation for minimized latency and seamless integration into existing enterprise data ecosystems.

Key attributes include:

Server customization: Tailored server configurations designed to meet specific organizational requirements.
Encryption: SSL-supported secure connections.
Elasticity: Independent scaling options for compute power and storage capacity.

Db2 Warehouse on Cloud

Previously known as "dashDB for Analytics," Db2 Warehouse on Cloud is a fully managed, scalable cloud data warehouse designed specifically for high-efficiency analytics and machine learning tasks.

Primary business benefits include:

Autonomous Operation: Driven by Db2’s self-optimizing engine featuring automated monitoring, failover capabilities, and DevOps integration.
Analytics Performance: Employs IBM BLU Acceleration technology such as columnar data storage, in-memory computation, compressed query processing, and data skipping to deliver exceptional performance for complex analytical workloads.
Concurrent Task Handling: Adaptive Workload Management dynamically assigns resources to ensure consistent performance during periods of heavy concurrency.
Built-in Machine Learning & Geospatial Support: Provides in-database machine learning functions (including Association Rules, k-means clustering, Regression) alongside Esri-compatible spatial analytics; also offers native Python support through Jupyter Notebooks.
Elastic Scalability: Allows separate scaling of compute and storage resources via web interface or API to align with evolving business requirements.
Robust Security: Implements encryption by default both in transit and at rest; incorporates advanced access controls such as data masking, row-level permissions, role-based security models, and comprehensive audit logging.
Polyglot Data Models: Supports relational (both columnar and row), geospatial, and NoSQL document formats (XML, JSON, BSON) with sophisticated compression techniques.
Multi-Cloud Availability: Deployable on IBM Cloud as well as AWS platforms.

Db2 BigSQL

In 2018, IBM rebranded its SQL engine as IBM Db2 Big SQL (Big SQL). This enterprise-grade, hybrid ANSI-compliant SQL engine operates on Hadoop, offering massively parallel processing (MPP) and enhanced data querying capabilities. It delivers benefits such as low latency, high throughput, robust security, broad SQL compatibility, and data federation.

Big SQL enables querying through a single database connection across multiple sources including HDFS, relational databases, NoSQL systems, object storage, and WebHDFS. It integrates with Hive, HBase, and Spark to provide unified access to data residing in Hadoop ecosystems and relational databases across cloud, on-premises, or hybrid infrastructures.

Designed for data scientists and analysts, Big SQL supports complex ad hoc queries with increased concurrency while requiring less hardware compared to alternative Hadoop SQL solutions. Its ANSI-compliant parser also facilitates processing of unstructured streaming data using new API interfaces.

As part of the IBM Common SQL Engine framework, Big SQL seamlessly interoperates with the Db2 family and IBM Integrated Analytics System. It constitutes a key element of IBM’s Hybrid Data Management Platform strategy focused on flexibility, portability, comprehensive data integration, and adaptable licensing options.

Db2 for IBM i

In 1994, IBM renamed the integrated relational database within OS/400 to DB2/400 to reflect feature parity with Db2 offerings on other platforms. Despite the Db2 branding, this product originated from the IBM System/38 integrated database architecture rather than from Db2 codebase. It is currently marketed as IBM Db2 for i.

Other Platforms

Db2 for Linux, UNIX and Windows (Db2 LUW)
Db2 for z/OS (mainframe), engineered for high availability, security, and performance in managing extensive data volumes
Db2 deployed on IBM Cloud

Editions

IBM provides four distinct Db2 editions: Community Edition, Starter Edition, Standard Edition, and Advanced Server Edition.

IBM Db2 Community Edition

The IBM Db2 Community Edition is a no-cost version available for download that delivers essential Db2 functionality. In version 11.5, it is restricted to 4 virtual processor cores and 16 GB of instance memory, without access to enterprise support or fix packs. Version 12.1 imposes a further limitation on server memory, capping it at 8 GB per server and restricting usage to non-production scenarios.

From version 11.5.1 onward, database size limits have been removed; earlier releases imposed a 100 GB maximum. The former free variant, IBM DB2 Express-C, supported up to 16 GB of RAM and two CPU cores.

Db2 Starter Edition

Launched with version 12.1, the Starter Edition is a commercial offering confined to 4 cores and 16 GB of memory. It does not include advanced capabilities such as native encryption, audit logging, federation, or Db2 pureScale.

IBM Db2 Standard Edition

The Standard Edition is offered under a perpetual licensing model for both production and non-production environments. It supports up to 16 processor cores and 128 GB of RAM with IBM support services. Licensing for production use is based on Virtual Processor Cores (whether physical or virtual), while non-production licenses are determined by the number of authorized users.

IBM Db2 Advanced Edition

Exclusively available through the IBM Hybrid Data Management Platform (HDMP), the Advanced Edition can be licensed either as a perpetual license or via monthly subscriptions for unlimited usage accompanied by premium IBM support. Licensing requires the acquisition of FlexPoints—flexible credits applicable across various Db2 products and IBM cloud offerings.

Technical information

Db2 administration can be performed using command-line tools or graphical user interfaces and supports both SQL and XQuery. It implements native XML storage by saving XML data in its original format rather than as relational data or CLOBs, which facilitates accelerated access through XQuery.

The platform offers APIs compatible with Rexx, PL/I, COBOL, RPG, Fortran, C++, C, Delphi, .NET CLI, Java, Python, Perl, PHP, Ruby, among others. Integration is supported with development environments such as Eclipse and Visual Studio.

pureQuery serves as IBM’s data access solution aimed at applications requiring data retrieval, supporting both Java and .NET platforms. It facilitates data interaction across databases and in-memory Java objects through tools, APIs, and runtime components included within IBM Data Studio Developer and IBM Data Studio pureQuery Runtime.

External links

Db2 - IBM Data for developers