Modernization Hub

Granularity

Enhanced Definition

In the mainframe context, **granularity** refers to the level of detail or the size of the smallest unit at which a resource, data, or process is managed, accessed, locked, or defined. It dictates the precision and scope of operations within z/OS systems, impacting performance, concurrency, and resource utilization.

Key Characteristics

    • Scope of Control: Defines the extent of an operation, such as locking a single row versus an entire table in a database, or managing a specific dataset versus an entire DASD volume.
    • Resource Unit Size: Determines the minimum allocable or manageable unit, like a 4KB page in DB2, a block on a DASD volume, or an individual record within a VSAM cluster.
    • Impact on Concurrency: Finer granularity (e.g., row-level locks) generally allows for higher concurrency in shared environments but can increase overhead. Coarser granularity (e.g., table-level locks) reduces overhead but limits concurrent access.
    • Performance Implications: The choice of granularity directly affects I/O operations, CPU usage, contention management, and the efficiency of data access and updates.
    • Data Detail Level: Pertains to how detailed data is stored or reported, ranging from individual transaction records to summarized aggregates for reporting or auditing.
    • Security Access Precision: Dictates how precisely security rules can be applied, e.g., at the dataset level, member level within a PDS/PDSE, or even field level within a database.

Use Cases

    • Database Locking (DB2, IMS): Specifying the lock size (e.g., ROW, PAGE, TABLESPACE in DB2, or SEGMENT, DBD in IMS) to balance concurrency and overhead for SQL or DL/I transactions.
    • Data Storage and Retrieval: Designing VSAM KSDS files where individual records are the units of access, or defining DB2 tables with specific ROWIDs for granular updates.
    • Resource Allocation: Managing memory in 4KB pages or allocating DASD space in tracks or cylinders for datasets.
    • Performance Monitoring: Collecting performance metrics at a very detailed level (e.g., individual CICS transaction response times, DB2 SQL statement execution details) versus aggregated hourly reports.
    • Security Access Control (RACF/ACF2/TSS): Defining access rules for specific datasets, members within a PDS/PDSE, or resources rather than broad categories, ensuring the principle of least privilege.
    • Backup and Recovery: Performing recovery operations at the level of a specific DB2 table, IMS database, or even a single VSAM cluster rather than an entire DASD volume.

Related Concepts

Granularity is intrinsically linked to concurrency control in database management systems like DB2 and IMS, where it directly influences the level of parallelism achievable by multiple users or applications. It significantly impacts performance tuning by

Related Vendors

ADPAC Corporation

5 products

Tone Software

14 products

IBM

646 products

Applied Software

7 products

Trax Softworks

3 products

Related Categories

Performance

171 products

Operating System

154 products

Tools and Utilities

519 products

Automation

222 products

Encryption

41 products