IDMS
| Company type | Software |
|---|---|
| Industry | Database management system |
| Predecessor | Integrated Data Store (IDS) |
| Founded | 1960s |
| Founder | Chemical Division of B.F. Goodrich |
| Fate | Acquired by Computer Associates in 1989; rebranded multiple times; currently owned by Broadcom |
Area served | Global |
Key people | Charles Bachman (leader of IDS development) |
| Products | Integrated Database Management System (IDMS) |
| Owner | Broadcom (since 2018) |
| Parent | Broadcom (after acquisition of CA Technologies) |
| Website | www |
The Integrated Database Management System (IDMS) is a network model (CODASYL) database management system created specifically for mainframes. Originally developed by the Chemical Division of B.F. Goodrich and marketed by Cullinane Database Systems (which later became Cullinet), the product has been under the ownership of Computer Associates (now CA Technologies) since 1989. Following the acquisition, it was rebranded as Advantage CA-IDMS and subsequently CA IDMS. In 2018, after Broadcom acquired CA Technologies, the name reverted back to IDMS.
History
IDMS originated from the groundbreaking database management system known as Integrated Data Store (IDS), which was created at General Electric under the leadership of Charles Bachman and launched in 1964. During the early 1960s, B.F. Goodrich’s Chemical Division undertook a rewrite of IDS using Intermediate System Language (ISL), a portable programming language designed to facilitate cross-platform compatibility.
Although it started as proprietary software, B.F. Goodrich transferred marketing rights to John Cullinane, enabling Cullinane Database Systems to commercialize IDMS. Due to royalty arrangements, supplementary products were invoiced separately, occasionally causing confusion for clients.
Initially deployed on GE-200 series and GE DATANET-30 machines, IDMS was subsequently adapted for IBM mainframe platforms (System/360, System/370, System/390, zSeries, System z9) as well as hardware from DEC and ICL. By the mid-1980s, approximately 2,500 licenses had been distributed to prominent organizations such as Strategic Air Command, Ford of Canada and Europe, Jaguar Cars, AXA/PPP, Tesco, Manulife, Bank of Canada, General Electric, Aetna, and BT UK.
A version tailored for DEC PDP-11 was sold under the name DBMS-11. In 1976, ICL obtained a source code license and ported IDMS to their 2900 series and 1900 range mainframes. ICL also independently developed upgraded variants like IDMSX, which found extensive use within UK government agencies—for example, Inland Revenue’s Pay-As-You-Earn system—some installations of which remained active as late as 2013.
The emergence of relational database management systems during the 1980s—propelled by advances in hardware performance and client-server architectures—led to growing adoption of relational databases due to their enhanced development efficiency despite early concerns over speed. In response, Cullinet introduced a relational front-end alongside productivity tools such as Automatic System Facility (ASF), utilizing IDMS’s Logical Record Facility (LRF) to facilitate application generation.
Nevertheless, these additions had limited enduring influence; users of IDMS predominantly appreciated its superior performance rather than relational capabilities. E. F. Codd’s distinction between authentic relational databases and network databases equipped with relational interfaces underscored this difference.
Following Computer Associates’ acquisition of Cullinet in 1989, Release 12.0 was launched in 1992–93 with comprehensive SQL support. CA Technologies continued to advance CA IDMS by incorporating features including TCP/IP connectivity, two phase commit, XML publishing capabilities, support for zIIP specialty processors, web-enabled access through CA IDMS Server and SQL Option, as well as GUI-based administration via the CA IDMS Visual DBA tool. Presently, CA-IDMS remains operational globally within mission-critical business environments, with numerous clients adopting web-enabled solutions through CA Technologies' Dual Database Strategy employing the CA-IDMS SQL Option.
Integrated Data Dictionary
One of the standout components of IDMS was its native Integrated data dictionary (IDD), designed to manage database definitions efficiently. The IDD functioned as an IDMS database itself, accessible to database administrators and users through the Data Dictionary Definition Language (DDDL). Additionally, it housed definitions and programming for other IDMS tools such as ADS/Online and IDMS-DC. Its adaptable architecture enabled organizations to develop custom definitions, frequently employed for internal documentation purposes.
Logical Data Model
IDMS utilized the CODASYL network model, which revolves around records and sets. Records were structured similarly to COBOL formats, containing fields of diverse types that could represent intricate data constructs like repeating groups.
Central to this model is the set, which embodies one-to-many relationships between records—one owner linked to multiple members. Unlike hierarchical models, a single record may participate in several sets, providing greater modeling versatility. Sets maintain an ordered sequence, allowing information to be conveyed through record order. Each record possesses a unique database key, commonly tied to its physical disk address. This close correlation between logical keys and physical storage facilitated rapid data retrieval but made data loading and restructuring operations resource-intensive.
Data access methods include direct retrieval via the database key, traversal through set relationships, or direct access using hashing known as CALC access. In CALC access, records sharing a hash value are connected to an owner record within each disk page’s header. Subsequent releases added BTree-like indexing structures to enhance access efficiency.
Storage
IDMS arranges databases into files divided into areas, which are further segmented into pages corresponding to disk blocks where records are stored. Database administrators allocate a fixed number of pages per area and specify how records are stored within them.
Specialized space allocation pages track free space but only monitor it when availability falls below 30%, minimizing input/output overhead.
There are four principal record storage techniques: Direct, Sequential, CALC, and VIA; Fujitsu/ICL’s IDMSX extends these with Page Direct and Random methods.
- Direct storage places records near a designated database key; applications receive the actual key after storage. - Sequential storage appends new records at the end of an area but is seldom employed. - CALC uses hashing for efficient placement and retrieval by linking records on pages through CALC sets owned by header records. This method typically achieves retrieval in approximately 1.1 I/O operations but encounters difficulties when primary keys change or areas expand, necessitating expensive reorganizations or performance-degrading workarounds. - VIA clusters member records close to their owners within sets, often on the same page, improving navigation speed; in IDMSX, VIA supports clustering across multiple areas.
IDMSX’s Page Direct mode targets specific database pages connected via CALC chains; Random mode assigns target pages based on CALC hashing seeded by keys or timestamps.
Sets are implemented as linked lists using database keys as pointers; designers decide whether to include owner or previous pointers, influencing navigation efficiency.
Later iterations of IDMS introduced indexing structures—record indexes facilitate secondary key lookups while set indexes enable keyed member retrieval. In IDMSX, Page Direct and Random placements generally operate alongside these indexes, which adhere to placement principles similar to Direct or CALC methods.
Release history (Computer Associates/CA period)
| Version | Date of release | Notable features |
|---|---|---|
| R12 | 1992 | Continuous 24-hour operation, logical and physical separation, catalog handling, deadlock resolution, centralized security framework, SQL support |
| R14 | 1999-01-09 | Utilization of Parallel Sysplex technology, multitasking capabilities |
| R15 | 2001-04-25 | Enhanced system performance, data sharing enhancements |
| R16 | 2004-04-13 | Implementation of two-phase commit, TCP/IP integration, Parallel Access Volumes utilization, XML publishing features |
| R17 | 2008-10-30 | Performance improvements, zIIP processor support, advanced auto-recovery mechanisms |
| R18 | 2011-06-02 | Further zIIP enhancements, automated system tuning, streamlined installation and upkeep processes |
| R19 |
User groups
IDMS is supported by multiple non-profit user organizations dedicated to CA IDMS and its related solutions, which include:
- CA IDMS (IUA EIUA) Global User Community
- AID – Anwendervereinigung IDMS Deutschland
- Australian IDMS User Group
- BeNeLux IDMS User Association
- Chicago IDMS Users Group
- Committee Quebec Users of IDMS
- Dallas CA-IDMS Users Forum
- European IDMS User Group
- Finnish IDMS User Group
- AUI – Association des Utilisateurs d'IDMS – France
- Rocky Mountain IDMS Users Group
- UK IDMS User Association
See also
- CA Datacom/DB
- IBM Information Management System
- ICL IDMS(X)