IBM Information Management System
| Company type | Software product |
|---|---|
| Industry | Information technology, Database management |
| Predecessor | ICS/DL/I |
| Founded | 1967 |
| Headquarters | Armonk , United States |
| Products | Hierarchical database, Transaction management system |
| Owner | IBM |
| Website | https://www.ibm.com/products/ims |
The IBM Information Management System (IMS) is a hierarchical database and information management solution designed to support transaction processing. Its development commenced in 1966 with the goal of managing the bill of materials for the Saturn V rocket utilized in the Apollo program. The initial release, named ICS/DL/I, was finalized in 1967 on the IBM System/360 Model 65 and deployed in August 1968.
IBM renamed the product to IMS/360 in 1969 and subsequently adapted it for newer platforms. By 1988, IBM reported that IMS was active at approximately 7,000 sites globally. The system remains widely implemented and continuously enhanced. Its highest sales volume occurred in 2003, 35 years post-launch, with usage by more than 95% of the Fortune 1000 companies.
DATE and DL/1
The origins of IMS trace back to a 1963 NASA contract aimed at managing change control for the Apollo command and service module at North American Rockwell’s facilities in Downey, California. IBM assigned Uri Berman to develop a solution on the IBM 7010 mainframe that would monitor part revisions.
Working alongside Rockwell engineer Pete Nordyke, they devised a system leveraging the then-novel hard disk, accessible through NAR’s engineering terminals. This database recorded part numbers alongside their latest revision data, enabling engineers to verify they used up-to-date components by querying relationships among parts.
This system was named DATE ("Disk Applications in a Teleprocessing Environment") and was architected with a clear distinction between physical disk storage and user interaction layers. The user interface component evolved into what became known as DL/1 ("Data Language/One"). DATE was installed on the IBM 7010 in 1965.
ICS
While DATE was operational, the IBM 7010 had become obsolete and was replaced by the System/360 series. North American Rockwell chose the IBM System/360 Model 65 as its successor. IBM proposed an enhanced DATE concept that separated data format definitions from the DL/1 query interface, allowing flexible data utilization through modifications to the data definition language. Although initially met with skepticism from Rockwell’s management, technical manager Robert Brown—an ex-IBM employee—supported this approach for its portability advantages.
The first formal specification emerged in 1966. The development team comprised twelve IBM personnel, ten from North American Rockwell, and three representatives from Caterpillar Tractor, which also intended to implement the system. This new solution was titled ICS/DL/I ("Information Control System and Data Language/Interface"). The inaugural version was completed in 1967 and installed at North American Rockwell on August 14, 1968, signaled by the initial "READY" prompt on an NAR IBM 2740 terminal.
IMS
In 1969, IBM introduced the product under the name IMS/360. It has undergone continuous evolution alongside advances in System/360 technology, extending into modern platforms such as z/OS and IBM zEnterprise System. By its twentieth anniversary in 1988, IMS was operational at roughly 7,000 sites; by late 1989, this number had grown to about 30,000, although revenue from Db2 had risen to rival that of IMS.
Vern Watts served as IMS’s principal architect for many years and had been involved with IMS since the 1960s. Joining IBM in 1956, he contributed at IBM’s Silicon Valley laboratories until his passing in 2009. Uri Berman, recognized as the "father of DL/1," received an IBM Outstanding Contribution Award along with a $100,000 prize for his significant role.
Database
The IBM Information Management System (IMS) employs a hierarchical database model, which differs fundamentally from IBM's subsequent relational database model utilized in Db2. Within IMS, information is arranged into segments—analogous to records or structs—each containing multiple fields and organized in a hierarchical manner to effectively handle complex data relationships. For instance, a customer database might have a root segment capturing details such as phone number, name, and age, with subordinate segments representing orders and further nested segments for individual order line items.
Full Function Databases
- Derived from Data Language Interface (DL/I) databases originally created for Apollo, full function databases support both primary and secondary indexing and are accessed through DL/I calls, which function similarly to SQL commands in IBM Db2 or Oracle.
- Access techniques include Hierarchical Direct (HDAM) and Hierarchical Indexed Direct (HIDAM), alongside other formats like Simple Hierarchical Indexed Sequential (SHISAM), Hierarchical Sequential (HSAM), and Hierarchical Indexed Sequential (HISAM).
- Data storage relies on VSAM, the native access method for z/OS systems, or OSAM, an IMS-specific approach that optimizes input/output operations tailored to IMS usage patterns, particularly benefitting from sequential access via OSAM Sequential Buffering.
Fast Path Databases
- Engineered to support extremely high transaction throughput, Fast Path databases encompass Data Entry Databases (DEDBs) and Main Storage Databases (MSDBs).
- DEDBs utilize direct (randomizer) access methods comparable to HDAM found in full-function databases, boosting performance through rapid access techniques and asynchronous write-after-commit capabilities. Since IMS Version 11, they incorporate the Unit of Work (UOW) concept to facilitate efficient online reorganization utilities.
High Availability Large Databases (HALDBs)
- Introduced with IMS Version 7, HALDBs extend the features of full function databases to improve availability and handle very large datasets.
- The capability for online reorganization was added in IMS Version 9, allowing continuous system operation while managing data volumes exceeding 40 terabytes.
Transaction Manager
IMS functions as a transaction manager (known as IMS TM or IMS DC), interfacing with users through VTAM, TCP/IP, 3270 terminals, Web interfaces, or other client applications. It executes business operations—such as processing banking withdrawals—while maintaining transactional state and ensuring precise recording within data repositories.
Operating on a messaging and queuing paradigm, IMS TM receives transactions from terminals or applications that are placed onto message queues. The IMS scheduler then triggers the relevant business application within a message processing region to handle the transaction by reading from or updating IMS or Db2 databases and ensuring correct recording. Responses are queued back and delivered to the originating terminal. IMS TM supports processing rates ranging from thousands up to tens of thousands of transactions per second; notably, IMS Version 13 demonstrated the ability to process 100,000 transactions per second on a single system.
Application
IMS TM offers a standardized environment for executing transactions. By the late 2000s, numerous Chinese financial institutions had adopted IMS to accommodate rapid expansion in their banking sectors.
Serving as a complement to IBM Db2—which was introduced in 1982 as a relational database—IMS generally provides superior speed for core transaction processing but can require more extensive programming effort for ancillary functions.
Relational data warehouses frequently augment IMS databases; for example, IMS efficiently manages primary ATM transaction processing due to its optimized performance tailored for specific operational workloads.