Implementation of IMS Databases

IMS DB supports multiple forms of enterprise databases, so that varied application requirements can be met by exploiting whichever database technology best suits the users' requirements.

The supported IMS database types are:

Full-function databases

Hierarchic databases that are accessed through Data Language I (DL/I) call language and can be processed by all six types of application programs: IFP, MPP, BMP, JMP, JBP, and batch.

Full-function databases can be accessed directly by record or sequentially, and by other sequences that are planned for when the database is designed. Most full-function databases are limited in size to 4 GB or 8 GB (depending on the access method) per data set. Data Set Groups, partitioning, or High Availability Large Databases (HALDBs) offer higher capacities.

Fast Path databases

Two types of databases designed to provide highly available data and fast processing for IMS applications. They can be processed by all types of application programs. The two types of Fast Path databases are:

Data entry databases (DEDBs)

A direct-access database that consists of one or more areas, with each area containing both root segments and dependent segments. DEDBs use a data structure that allows them to be used for both hierarchic processing and journaling. The database is accessed by using VSAM's Media Manager.

DEDBs are particularly suited for use where large databases or very low processing costs are required, or when particularly high data availability or very high performance is required.

Main storage databases (MSDBs)

MSDB functionality has been superseded by the DEDB Virtual Storage Option (VSO), so this book does not describe MSDBs and you should not to use them.

IMS uses a hierarchical model for its database, which is described in more detail in Chapter 7, "Overview of the IMS Hierarchical Database Model," on page 67. The data stored in the IMS databases is organized using a number of internal IMS access methods. Each of these access methods is tailored for certain types of access to the database. The choice of the appropriate access method is discussed in detail in Chapter 8, "Implementing the IMS Hierarchical Database Model," on page 83.

No single database technology is the best option for all applications, even though industry trends might suggest that an organization standardize on only one database type (for example, a relational database). However, limiting your enterprise to only relational databases would preclude the consideration of other technologies that might result in significant savings in processing time or application development costs far in excess, for example, of the small additional cost of introducing DEDBs to your organization.

Compared to IMS DB, DB2 UDB for z/OS provides well for unstructured or unplanned access to data and therefore provides flexibility in the support of future application requirements. However, any IMS database usually has a significantly lower processing cost than a DB2 UDB for z/OS database.

The IMS access methods are the application's view of how the data is stored. IMS actually uses the operating system access methods to physically store data on disk storage. The software access methods that IMS uses are:

• VSAM (Virtual Storage Access Method): A z/OS access method.

• OSAM (Overflow Sequential Access Method): An IMS data management access method that combines selected characteristics of z/OS BSAM (Basic Sequential Access Method) and BDAM (Basic Direct Access Method).

Full-Function Databases

Full-function databases are designed to support most types of database requirements and can be used in a wide variety of applications. Most IMS applications make use of full-function databases unless there are specific requirements for DEDBs. The major characteristics of full-function databases are:

• Small or large databases

• Access to records through unique or non-unique keys

• Many types of segments (up to 15 levels allowed)

• Records can be stored in key sequence, but this is not required

One function associated with full-function databases is called data set groups. With data set groups, you can put some types of segments in a database record in data sets other than the primary data set without destroying the hierarchic sequence of segments in a database record. You might use data set groups to accommodate the differing needs of your applications. By using data set groups, you can give an application program fast access to the segments in which it is interested. The application program simply bypasses the data sets that contain unnecessary segments. You can define up to 10 data set groups for a single full-function database.

The full-function databases that were created in IMS Version 1 through IMS Version 6 were limited in size: the maximum data set size for VSAM is 4 GB and for OSAM is 8 GB. IMS Version 7 introduced High Availability Databases (HALDBs) to address this size limit. HALDB allows full-function databases to grow much larger. A HALDB is a partitioned full-function database. Partitioning a database allows the use of smaller data sets that are easier to manage. Multiple partitions decrease the amount of unavailable data if a partition fails or is taken offline. HALDBs are implemented with OSAM or VSAM, but always with a maximum 4 GB data set size.

HALDB allows the grouping of full-function database records into sets of partitions that are treated as a single database while permitting functions to be performed independently for each partition. Each HALDB partition has the same capacity limit as a non-HALDB database: each partition can consist of up to 10 data sets.

You can increase the amount of data that is stored in a single partition by using data set groups. The logical design of a database applies for the entire HALDB database, not independently for each partition. The choice to use data set groups, which allow the storage of selected segments into separate nominated data sets, is a logical database design decision. Each HALDB partition has the same logical design because each partition has the same architecture for data set groups. By using data set groups in partitions, a large amount of data can be contained in a single partition. HALDBs can contain up to 1001 partitions.

Related Reading: For more information about HALDBs, see "PHDAM and PHIDAM Access Methods" on page 97.

Data Entry Databases (DEDBs)

DEDBs support intensive IMS database requirements, particularly in the banking industry, for:

• Large databases that contain millions of records and extend well beyond the original 4 GB database limits for full-function databases

• Access to each database record by a key field

• Lower processing costs for each database record and update than are required for full-function databases

• The capability to support higher transaction workloads than full-function databases can sustain, while maintaining per-transaction cost advantages

• Improved availability, with reduced requirements for database outage, especially for database maintenance activities such as database reorganizations

• Lower processing costs for particular types of processing, where data is inserted online and retrieved in batches for further processing, and eventually deleted in batches

• The possibility of eliminating transaction-related I/O from database processing

Application programming for DEDBs is little different from that for full-function databases.