Appendix H: Glossary

This glossary defines the key terms used throughout this textbook. Each entry includes a brief definition and the chapter(s) where the term is introduced or discussed in depth. Terms are organized alphabetically. Where a term has different meanings in different contexts (e.g., "region" means something different in JCL than in CICS), all relevant definitions are provided.

A

Above the bar. The 64-bit virtual storage area above the 2 GB boundary (the "bar") in a z/OS address space. COBOL programs can use this storage via MEMLIMIT and Language Environment 64-bit services. Chapters 2, 3.

Above the line. The virtual storage area between 16 MB and 2 GB in a z/OS address space. Most modern COBOL programs execute here (AMODE 31). Chapters 1, 2.

Access path. The strategy DB2 chooses to retrieve data for a SQL statement — tablespace scan, index scan, index-only access, list prefetch, or multi-index access. Determined by the cost-based optimizer using catalog statistics. Chapters 6, 11.

ACCESSTYPE. A column in DB2's PLAN_TABLE that indicates the access method chosen for a table: I (index), R (tablespace scan), N (workfile), M (multi-index). Chapter 11.

ACH (Automated Clearing House). An electronic payment network for financial transactions in the United States. CNB processes millions of ACH transactions daily through batch COBOL programs. Chapters 22, 23, 38.

Address space. A z/OS construct providing each task (batch job, CICS region, DB2 subsystem) with its own isolated virtual storage environment. Hardware-enforced isolation prevents one address space from corrupting another. Chapter 1.

AMODE (Addressing Mode). Specifies whether a program uses 24-bit (AMODE 24), 31-bit (AMODE 31), or 64-bit (AMODE 64) addresses. Most production COBOL programs run AMODE 31. Chapters 2, 3.

AOR (Application Owning Region). A CICS region type that executes application programs. In a multi-region topology, AORs are separated from terminal-owning and file-owning regions for scalability and failure isolation. Chapters 13, 17.

APF (Authorized Program Facility). A z/OS mechanism that designates load libraries as authorized, allowing programs in those libraries to issue privileged system calls. Chapter 1.

APREUSE. A DB2 BIND/REBIND option that instructs DB2 to reuse the existing access path unless the new path is significantly better. Used to prevent access path regressions after RUNSTATS. Values: WARN, ERROR. Chapter 6.

AT-TLS (Application Transparent Transport Layer Security). A z/OS Communication Server feature that provides TLS encryption at the TCP/IP stack level, transparent to the application. Chapter 28.

Audit trail. A chronological record of system activities sufficient to reconstruct, review, and examine the sequence of events surrounding or leading to each activity. Required by PCI-DSS, HIPAA, SOX. Chapters 7, 28.

B

Backout queue. An MQ queue where poison messages are routed after exceeding the backout threshold (BOTHRESH). Prevents a single bad message from blocking the entire queue. Chapter 19.

Batch window. The scheduled time period (typically overnight) when online systems are quiesced or reduced and batch processing runs. The batch window is a scheduling problem, not just a performance problem. Chapters 23, 25, 26.

Below the line. The virtual storage area below 16 MB in a z/OS address space. Constrained and valuable — only programs requiring AMODE 24 compatibility use this area. Chapters 1, 2.

BIND (DB2). The process of converting a DB2 application program's SQL statements (in a DBRM) into an executable access path (a package or plan). BIND is when the optimizer makes its access path decisions. Chapters 6, 12.

BMS (Basic Mapping Support). A CICS facility for formatting 3270 terminal screens. BMS maps define the screen layout; the COBOL program populates the data. Chapter 13.

BOTHRESH (Backout Threshold). An MQ queue attribute specifying the maximum number of times a message can be backed out before being routed to the backout queue. Chapter 19.

Buffer pool (DB2). An area of virtual storage used to cache DB2 data and index pages. Properly sized buffer pools reduce physical I/O. DB2 supports multiple buffer pools (BP0–BP49, BP8K0–BP8K9, BP16K0–BP16K9, BP32K, BP32K1–BP32K9). Chapters 6, 9, 11.

BUFNO. A JCL DCB parameter specifying the number of I/O buffers for a dataset. Increasing BUFNO improves sequential processing performance by enabling look-ahead reads. Chapter 26.

C

Capacity planning. The process of forecasting future resource requirements (CPU, storage, I/O, network) based on current consumption trends and projected business growth. Chapter 29.

CARDF (Cardinality). A DB2 catalog statistic in SYSIBM.SYSTABLES indicating the number of rows in a table. Used by the optimizer for filter factor calculations. Chapter 6.

CDE (Cardholder Data Environment). The PCI-DSS term for the systems, people, and processes that store, process, or transmit cardholder data. Defining the CDE boundary is critical for scope management. Chapter 28.

CEEDOPT. The Language Environment default options table linked into a COBOL program's load module. Overrides system-level LE defaults. Chapter 3.

CEEUOPT. The Language Environment user options table that overrides CEEDOPT at execution time. Used to change LE behavior without relinking. Chapter 3.

CEE3ABD. A Language Environment abend macro that produces a formatted dump (CEE3DMP) before terminating the enclave. Critical for post-mortem diagnosis of production abends. Chapter 3.

Channel (CICS). A named, typed data-passing mechanism in CICS that replaces the 32 KB COMMAREA limitation. Channels contain one or more named containers of arbitrary size. Chapters 15, 21.

Channel (MQ). A communication link between two MQ queue managers. Sender-receiver channel pairs transmit messages between queue managers. Channels can be triggered or permanently running. Chapter 19.

Checkpoint/restart. A technique where a batch program periodically saves its processing state (checkpoint) so that after an abend, it can resume processing from the last checkpoint rather than restarting from the beginning. Chapter 24.

CICS (Customer Information Control System). IBM's transaction processing monitor for z/OS. Manages online transaction processing, including terminal management, program execution, resource management, and recovery. Chapters 13–18.

CICSPlex SM (CICSPlex System Manager). IBM's management product for multi-region CICS environments. Provides workload management, topology management, operations management, and resource monitoring across a CICSplex. Chapter 13.

CLUSTERRATIO. A DB2 catalog statistic indicating the percentage of rows in a table that are physically stored in the same order as the clustering index. Low CLUSTERRATIO degrades index scan performance and triggers REORG. Chapters 6, 9.

CMAS (CICSPlex SM Address Space). The address space that runs CICSPlex SM management functions. Typically deployed in an HA pair. Chapter 13.

COBOL (Common Business-Oriented Language). A programming language designed for business data processing, first specified in 1959. Enterprise COBOL for z/OS is IBM's current implementation. All chapters.

COLCARDF (Column Cardinality). A DB2 catalog statistic in SYSIBM.SYSCOLUMNS indicating the number of distinct values in a column. Used by the optimizer to calculate filter factors. Chapter 6.

COMMAREA (Communication Area). A data area passed between CICS programs or between transaction steps (pseudo-conversational). Limited to 32,763 bytes. Largely superseded by channels and containers. Chapters 13, 15.

COMMIT. The operation that makes all database changes within the current unit of work permanent and releases locks. In COBOL: EXEC SQL COMMIT (batch) or EXEC CICS SYNCPOINT (CICS). Chapters 8, 24.

Commit frequency. The number of records or operations processed between COMMIT points in a batch program. Balances lock duration, restart granularity, and log volume. Chapters 8, 24.

Container (CICS). A named data area within a CICS channel. Containers can hold binary or character data of any size. Chapter 15.

Copybook. A COBOL source member included via the COPY statement. Used to share data definitions (record layouts, SQL host variables, MQ structures) across programs. All chapters.

Cost-based optimizer. The DB2 component that selects access paths by estimating the CPU and I/O cost of alternative strategies using catalog statistics. Chapter 6.

Coupling Facility (CF). Specialized hardware in a Parallel Sysplex that provides high-speed, shared-access data structures (lock lists, cache structures, queue structures) for cross-LPAR coordination. Chapters 1, 8, 13, 30.

CPACF (CP Assist for Cryptographic Functions). Hardware acceleration for cryptographic operations on z/OS. Makes encryption effectively free in terms of CPU cost. Chapter 28.

Critical path. In batch scheduling, the longest sequence of dependent jobs through the batch window DAG. The critical path determines the minimum possible batch window duration. Chapter 23.

Cross-memory communication. A z/OS mechanism allowing one address space to access data or call routines in another address space. Used by DB2 (COBOL program calls DB2 via cross-memory), CICS MRO, and other subsystem interactions. Chapters 1, 13.

CS (Cursor Stability). A DB2 isolation level that holds read locks only while the cursor is positioned on a row. Provides a good balance between concurrency and consistency for most online applications. Chapter 8.

CSA/ECSA (Common Service Area / Extended CSA). z/OS virtual storage areas accessible to all address spaces on the LPAR. Used for inter-address-space communication. CSA is below the line; ECSA is above. Chapter 2.

CTE (Common Table Expression). A temporary named result set defined within a SQL statement using the WITH clause. Recursive CTEs enable hierarchical data processing. Chapter 7.

D

DAG (Directed Acyclic Graph). A graph with directed edges and no cycles. Used to model batch job dependencies — jobs are nodes, dependencies are edges. Chapter 23.

Data sharing. A DB2 configuration where multiple DB2 subsystems on different LPARs share access to the same data via the Coupling Facility. Enables cross-LPAR workload distribution and HA. Chapters 1, 8, 30.

Datagram. An MQ message pattern where the sender sends a message and does not expect a reply. Also called "fire and forget." Chapter 19.

DBRM (Database Request Module). The output of the DB2 precompiler containing a program's SQL statements. Input to the BIND process. Chapter 12.

Dead letter queue (DLQ). An MQ queue where undeliverable messages are placed. Every queue manager should have a DLQ, and every production environment should have a DLQ handler program. Chapter 19.

Deadlock. A situation where two or more programs are each waiting for a lock held by the other, creating a circular wait. DB2 detects deadlocks and rolls back one participant (the "victim"). Chapter 8.

DFSMS (Data Facility Storage Management Subsystem). The z/OS component that manages dataset storage, including allocation, migration, backup, and space management. Includes DFSMSdfp, DFSMShsm, DFSMSrmm. Chapter 4.

Dispatcher. The z/OS component that allocates CPU time to ready tasks based on dispatching priority. WLM influences dispatcher decisions through service class goals. Chapters 1, 5.

DR (Disaster Recovery). The set of policies, procedures, and technologies that enable recovery of critical systems after a catastrophic failure. Chapter 30.

DSN_STATEMNT_TABLE. A DB2 table populated during EXPLAIN that contains cost estimates and performance predictions for SQL statements. Chapter 11.

Dynamic SQL. SQL statements that are constructed and prepared at runtime (PREPARE/EXECUTE) rather than being embedded in the program at compile time. Requires careful security handling to prevent SQL injection. Chapter 12.

E

EBCDIC (Extended Binary Coded Decimal Interchange Code). The character encoding used natively on z/OS. COBOL programs on z/OS process EBCDIC data; conversion to ASCII/UTF-8 is required for distributed system integration. Chapters 12, 22.

Enclave. A Language Environment execution scope. A COBOL program runs within an enclave; LE runtime options, storage management, and condition handling operate at the enclave level. Chapter 3.

Entry thread (DB2). A DB2 thread dedicated to a specific CICS transaction or plan. Provides guaranteed access and avoids pool thread contention for high-priority transactions. Chapter 1.

ESDS (Entry-Sequenced Dataset). A VSAM dataset organization where records are stored in the order they were written. No key-based access. Used for logs and journals. Chapter 4.

Event-driven architecture. An architectural pattern where system components communicate through events (asynchronous messages) rather than direct calls. MQ and CICS event processing support this pattern on z/OS. Chapter 20.

EXPLAIN. A DB2 facility that populates diagnostic tables (PLAN_TABLE, DSN_STATEMNT_TABLE, DSN_FILTER_TABLE) with the optimizer's chosen access path for SQL statements. Essential for performance diagnosis. Chapters 6, 11.

F

FAIL-IF-QUIESCING. An MQ option that causes an MQI call to fail immediately if the queue manager is shutting down. Must be set on all production MQ calls to enable graceful quiesce. Chapter 19.

Filter factor. The estimated fraction of rows that will satisfy a SQL predicate. Calculated by the DB2 optimizer using catalog statistics. A filter factor of 0.01 means 1% of rows are expected to qualify. Chapter 6.

FOR (File Owning Region). A CICS region type that owns VSAM files. AORs access files on FORs via function shipping (MRO). Provides file-level isolation and centralized file management. Chapter 13.

FREESPACE. A VSAM parameter specifying the percentage of free space to leave in control intervals and control areas during initial load or reorganization. Accommodates future inserts. Chapter 4.

Function shipping. A CICS mechanism where a program in one region (typically an AOR) issues a file or queue request that is automatically shipped to and executed in another region (typically a FOR). Transparent to the application program. Chapter 13.

G

GDG (Generation Data Group). A collection of chronologically related datasets (generations) sharing a common base name. Each batch run creates a new generation (+1); previous generations are referenced by relative number (0, -1, -2). Chapters 4, 23.

GDPS (Geographically Dispersed Parallel Sysplex). An IBM automated solution for disaster recovery that manages data replication and failover between primary and secondary sites. Chapter 30.

Goal mode. The WLM operating mode where resource allocation is driven by service-level goals (response time, velocity) rather than fixed priorities. The standard mode for modern z/OS installations. Chapter 5.

H

HANDLE ABEND. A CICS command that registers an abend exit routine. When the program abends, control transfers to the registered routine instead of immediately terminating the task. Chapter 18.

HANDLE CONDITION. A CICS command that registers condition handling routines. Older approach; RESP/RESP2 checking on individual commands is preferred in modern CICS programming. Chapter 18.

Heap. In Language Environment, a storage area for dynamically allocated data. Managed by LE heap management services. Heap fragmentation can cause S878 abends. Chapter 3.

HIPAA (Health Insurance Portability and Accountability Act). U.S. federal law establishing standards for protecting health information. Pinnacle Health's security architecture is driven by HIPAA requirements. Chapter 28.

Host variable. A COBOL data item used in embedded SQL to pass values between the COBOL program and DB2. Declared in the WORKING-STORAGE SECTION and referenced in SQL with a colon prefix (:WS-ACCT-ID). All DB2 chapters.

I

ICF (Integrated Catalog Facility). The z/OS catalog system that maps dataset names to physical storage locations (volumes). Master catalogs and user catalogs form a hierarchical name resolution structure. Chapter 4.

ICSF (Integrated Cryptographic Service Facility). A z/OS component providing cryptographic services including key management. Works with CPACF hardware and RACF CSFKEYS class. Chapter 28.

IFCID (Instrumentation Facility Component Identifier). A DB2 trace event identifier. IFCID 0172 reports deadlock information; IFCID 0196 reports lock timeout information; IFCID 0003 reports accounting data. Chapters 8, 11.

Image copy. A DB2 backup of a tablespace or partition. Full image copies capture the entire object; incremental copies capture only changed pages since the last full copy. Chapter 9.

IMS (Information Management System). IBM's hierarchical database and transaction manager. Federal Benefits Administration's 40-year codebase is heavily IMS-based. Chapters 1, 32.

Index-only access. A DB2 access path where all columns needed by the query exist in the index, eliminating the need to read the base table data pages. Indicated by INDEXONLY = 'Y' in PLAN_TABLE. Chapter 6.

Initiator. A z/OS address space that executes batch jobs. Each initiator can run one job at a time. The number of active initiators limits batch concurrency. Chapters 1, 23.

IPIC (IP Interconnectivity). A CICS intercommunication protocol based on TCP/IP. Supports both MRO-like local and ISC-like remote communication. Chapters 13, 14.

IRLM (Internal Resource Lock Manager). The z/OS subsystem that manages DB2 locks. Runs in its own address space and maintains lock tables in storage. Chapter 8.

ISC (Inter-System Communication). CICS communication between regions on different LPARs or different physical systems. Uses IPIC (TCP/IP) or older LU 6.2 SNA protocols. Chapter 13.

Isolation level. A DB2 setting that controls how a program's SQL statements interact with concurrent changes by other programs. Levels: UR (uncommitted read), CS (cursor stability), RS (read stability), RR (repeatable read). Chapter 8.

J

JCL (Job Control Language). The z/OS language for defining batch jobs, including program execution, dataset allocation, and job step sequencing. Chapters 1, 23, 26.

JES2 (Job Entry Subsystem 2). The z/OS subsystem that receives batch jobs, queues them for execution, manages job output (SYSOUT), and tracks job status. Chapter 1.

JSON (JavaScript Object Notation). A lightweight data interchange format. CICS supports JSON transformation for REST API integration with COBOL programs. Chapters 14, 21.

K

KSDS (Key-Sequenced Dataset). A VSAM dataset organization where records are stored in key sequence. Supports both sequential and keyed access. The most common VSAM type for COBOL applications. Chapter 4.

L

Language Environment (LE). The common runtime environment for z/OS high-level languages (COBOL, PL/I, C, Fortran). Provides initialization, storage management, condition handling, and termination services. Chapter 3.

List prefetch. A DB2 access method that collects RIDs (Record Identifiers) from an index, sorts them by page number, and then fetches the data pages in sequential order. Indicated by PREFETCH = 'L' in PLAN_TABLE. Chapters 6, 11.

Lock avoidance. A DB2 technique that avoids acquiring a lock on a page if the page has not been modified since the last commit point. Reduces lock overhead for read-mostly workloads. Chapter 8.

Lock escalation. The process where DB2 promotes many fine-grained locks (page or row) to a single coarse-grained lock (tablespace). Triggered when lock count exceeds LOCKMAX. Improves lock management efficiency but reduces concurrency. Chapter 8.

LOCKMAX. A DB2 tablespace attribute specifying the maximum number of locks that can be held on the object before escalation occurs. Chapter 8.

LOCKSIZE. A DB2 tablespace attribute specifying the default granularity of locks: ANY (DB2 chooses), PAGE, ROW, TABLE, TABLESPACE. Chapter 8.

LPAR (Logical Partition). A subdivision of a z/OS mainframe's physical resources (CPUs, memory, I/O) into an independently operating z/OS image. Multiple LPARs share one physical machine but are logically isolated. Chapter 1.

M

MATCHCOLS. A column in DB2's PLAN_TABLE indicating the number of index key columns that match predicates in the WHERE clause. Higher is generally better. Chapters 6, 11.

MAXTASK. A CICS system initialization parameter that limits the maximum number of concurrent tasks in a CICS region. Prevents resource exhaustion. Chapter 17.

MEMLIMIT. A JCL or system parameter that sets the upper limit for storage above the 2 GB bar available to a program. Chapter 2.

MERGE statement. A DB2 SQL statement that combines INSERT and UPDATE operations into a single statement. "Upsert" functionality. Chapter 7.

Metro Mirror (PPRC). IBM's synchronous disk replication technology. Provides zero data loss (RPO=0) but requires the DR site to be within a limited distance of the primary site. Chapter 30.

MQ (IBM MQ, formerly WebSphere MQ, MQSeries). IBM's message-oriented middleware providing asynchronous, reliable message passing between applications. Chapters 19, 20.

MQGET. The MQ API call to retrieve a message from a queue. Options control wait behavior, message selection, and conversion. Chapter 19.

MQMD (MQ Message Descriptor). The metadata header accompanying every MQ message. Contains message ID, correlation ID, persistence, expiry, format, priority, and other attributes. Chapter 19.

MQPUT. The MQ API call to place a message on a queue. Options control syncpoint participation, persistence, and message properties. Chapter 19.

MRO (Multi-Region Operation). A CICS inter-region communication mechanism for regions on the same LPAR. Uses cross-memory services rather than TCP/IP. Faster than ISC. Chapter 13.

MSU (Million Service Units). The unit of measurement for mainframe processing capacity. MSU consumption drives z/OS software pricing. Chapter 29.

Multi-row FETCH. A DB2 feature allowing a single FETCH statement to retrieve multiple rows into host variable arrays. Dramatically reduces cross-memory call overhead for high-volume reads. Chapter 7.

N

Nested loop join. A DB2 join method where for each row of the outer table, DB2 searches the inner table (typically via index). Efficient when the outer table has few qualifying rows. Chapters 6, 11.

O

OLAP functions. SQL functions for analytical processing within a query: ROW_NUMBER, RANK, DENSE_RANK, LAG, LEAD, running aggregates (SUM OVER, COUNT OVER). Chapter 7.

Optimistic locking. A concurrency control technique where the program reads a row without holding a lock, allows user "think time," then verifies that the row hasn't changed before updating. Eliminates lock contention during think time. Chapter 8.

P

Package (DB2). The executable form of a single DB2 application program's SQL statements after BIND processing. Contains access paths, host variable definitions, and authorization information. Chapter 12.

Parallel Sysplex. A cluster of up to 32 z/OS LPARs connected via Coupling Facilities, providing workload balancing, data sharing, and high availability. Chapters 1, 30.

Partition-parallel processing. A batch architecture pattern where a table partitioned by key range is processed by multiple parallel jobs, each handling one or more partitions. Chapter 25.

PCI-DSS (Payment Card Industry Data Security Standard). A security standard for organizations that handle credit card data. CNB's security architecture is built around PCI-DSS requirements. Chapter 28.

Performance index (PI). A WLM metric indicating how well a service class is meeting its goal. PI=1.0 means exactly meeting the goal; PI<1.0 is exceeding the goal; PI>1.0 is missing the goal. Chapter 5.

PLAN_TABLE. The primary DB2 EXPLAIN output table containing access path details: access type, index used, matching columns, join method, sort operations, prefetch strategy. Chapters 6, 11.

PLANMGMT. A DB2 BIND option that preserves previous access paths (PREVIOUS and ORIGINAL versions) so they can be restored if a new access path performs poorly. Values: OFF, BASIC, EXTENDED. Chapter 6.

Poison message. An MQ message that cannot be processed successfully and causes the consuming program to abend repeatedly. Without backout handling, a poison message blocks the entire queue. Chapter 19.

Pool thread (DB2). A shared DB2 thread available to any CICS transaction or batch program. Less expensive than entry threads but subject to contention at high volumes. Chapter 1.

PPRC (Peer-to-Peer Remote Copy). IBM's synchronous disk mirroring technology between primary and secondary storage subsystems. Foundation for Metro Mirror. Chapter 30.

Prefetch. A DB2 I/O optimization that reads multiple pages ahead of the current position. Types: sequential (S), list (L), dynamic. Controlled by the optimizer and indicated in PLAN_TABLE. Chapters 6, 11.

Pseudo-conversational. A CICS programming technique where the program terminates between user interactions, freeing resources during "think time." The program is restarted when the user presses Enter. Chapters 13, 15.

Pub/sub (Publish/Subscribe). An MQ messaging pattern where publishers send messages to topics and subscribers receive messages from topics they've subscribed to. Decouples publishers from subscribers. Chapter 20.

Q

Queue manager. The MQ component that manages queues, channels, and message storage on a single system. Identified by a unique name (e.g., CNBPQM01). Chapter 19.

Queue sharing group. An MQ configuration where multiple queue managers on different LPARs share access to the same queues via the Coupling Facility. Provides MQ high availability. Chapter 19.

R

RACF (Resource Access Control Facility). IBM's security product for z/OS. Authenticates users, authorizes resource access, and logs security events. The foundation of all z/OS security. Chapter 28.

Recursive CTE. A common table expression that references itself, enabling iterative processing of hierarchical data (org charts, bill-of-materials, graph traversal) within a single SQL statement. Chapter 7.

REGION parameter. A JCL parameter specifying the maximum amount of virtual storage available to a batch job step below the 2 GB bar. Chapter 2.

REORG. A DB2 utility that physically reorganizes a tablespace, restoring clustering order and reclaiming free space. Essential for maintaining query performance. Chapters 9, 26.

REORP (REORG Pending). A DB2 restrictive state indicating that a tablespace requires reorganization before normal access is allowed. Caused by certain schema alterations. Chapter 9.

Request/reply. An MQ messaging pattern where the sender puts a request message and expects a reply message in return. The MQMD-CORRELID links the reply to the original request. Chapter 19.

RESP/RESP2. CICS command response fields. RESP contains the primary response (NORMAL, ERROR, NOTFND, etc.); RESP2 contains additional diagnostic information. The preferred error handling mechanism in modern CICS COBOL. Chapter 18.

RMODE (Residency Mode). Specifies where in virtual storage a program's executable code must reside: RMODE 24 (below 16 MB), RMODE 31 or ANY (below 2 GB). Chapters 2, 3.

RPO (Recovery Point Objective). The maximum acceptable amount of data loss measured in time. RPO=0 means zero data loss; RPO=1hr means up to 1 hour of data may be lost. Chapter 30.

RR (Repeatable Read). A DB2 isolation level that holds read locks on all accessed pages/rows until COMMIT. Prevents phantoms and non-repeatable reads but maximizes lock contention. Chapter 8.

RS (Read Stability). A DB2 isolation level that holds read locks only on qualifying rows until COMMIT. Prevents non-repeatable reads on qualifying data but allows phantoms. Chapter 8.

RTO (Recovery Time Objective). The maximum acceptable duration between a disaster and restoration of service. Chapter 30.

RUNSTATS. A DB2 utility that collects statistical information about tables, indexes, and tablespaces and stores it in the DB2 catalog. The optimizer depends on current RUNSTATS for good access path decisions. Chapters 6, 9.

S

SAF (System Authorization Facility). A z/OS interface that routes all security decisions to the installed security product (typically RACF). Provides a single checkpoint for all resource access. Chapter 28.

Sequential prefetch. A DB2 I/O optimization that reads up to 32 consecutive pages in a single I/O operation. Used for tablespace scans and clustered index scans. Chapters 6, 26.

Service class. A WLM construct that groups related work and defines performance goals. Each service class has one or more periods with specific goals (response time or velocity). Chapter 5.

SHAREOPTIONS. A VSAM parameter controlling how multiple programs can concurrently access a dataset. Options range from single-user (SHAREOPTIONS 1) to full cross-system sharing (SHAREOPTIONS 4). Chapter 4.

Shared temporary storage. A CICS facility where temporary storage queues are stored in the Coupling Facility, accessible from any CICS region in the Sysplex. Eliminates inter-transaction affinity to specific AORs. Chapter 13.

SKIP LOCKED DATA. A DB2 feature that causes a query to skip rows that are locked by other transactions rather than waiting. Useful for "next available item" patterns. Chapter 8.

SMF (System Management Facilities). A z/OS facility that records system activity data in SMF records. Used for accounting, performance analysis, capacity planning, and security auditing. Chapters 27, 28, 29.

SMS (Storage Management Subsystem). See DFSMS. The system of rules (data class, storage class, management class) that automates dataset storage management. Chapter 4.

SORT. The z/OS utility (DFSORT or SyncSort) that sorts, merges, and copies datasets. Heavily used in batch processing. SORT optimization is a key batch performance technique. Chapter 26.

SOX (Sarbanes-Oxley Act). U.S. federal law requiring financial reporting controls and audit trails. Affects CNB's and SecureFirst's security and audit architecture. Chapter 28.

SQL injection. A security vulnerability where an attacker manipulates dynamic SQL by inserting malicious input. Prevented in COBOL by using host variables and parameter markers instead of string concatenation. Chapters 12, 28.

SQA/ESQA (System Queue Area / Extended SQA). z/OS virtual storage areas reserved for system use. Not available to application programs. Excessive SQA consumption can cause system-wide problems. Chapter 2.

Staged commit. A batch processing pattern where work is accumulated in memory and committed in groups (e.g., every 1,000 records) rather than after every record or only at end of job. Chapter 12.

Strangler fig pattern. A modernization pattern where new functionality is incrementally built alongside the legacy system, gradually taking over until the legacy system can be decommissioned. Named after the strangler fig tree. Chapter 33.

Stored procedure. A program registered in DB2 that can be called via the SQL CALL statement. COBOL stored procedures run in WLM-managed address spaces and can return result sets. Chapter 10.

SVC (Supervisor Call). A hardware instruction that transfers control from a user program to z/OS. Used for system services (dataset I/O, storage management, security checks). Chapter 1.

SYNCPOINT. A CICS command that commits the current unit of work (equivalent to SQL COMMIT in CICS). All DB2 updates, MQ messages, and VSAM changes are committed atomically. Chapters 8, 18, 19.

Sysplex. A general term for a cluster of z/OS systems. See Parallel Sysplex. Chapter 1.

T

Tablespace. The DB2 physical storage structure that contains one or more tables. Types include segmented, partitioned, universal (partition-by-growth, partition-by-range). Chapters 6, 9.

Tablespace scan. A DB2 access method that reads every page in a tablespace sequentially. Efficient for queries that access a large percentage of rows; inefficient for selective queries. Chapter 6.

Temporal tables. DB2 tables that automatically maintain historical versions of rows. Types: system-time (maintains history of all changes), business-time (models business-effective periods), bitemporal (both). Chapter 7.

Thread (DB2). A DB2 connection from an application program. Types: entry (dedicated), pool (shared). Threads consume DB2 resources and are limited by configuration. Chapters 1, 12.

TOR (Terminal Owning Region). A CICS region type that manages terminal connections (3270, TCP/IP) and routes transactions to AORs. Chapter 13.

Transaction affinity. A constraint that requires consecutive steps of a pseudo-conversational transaction to execute in the same AOR. Usually caused by storing state in region-local resources (TS queues, GETMAIN storage). Undesirable because it prevents workload balancing. Chapter 13.

Transaction routing. The CICS mechanism that directs a transaction from the TOR where it's entered to the AOR where it will execute. Can be static (configured) or dynamic (CICSPlex SM workload management). Chapter 13.

Transmission queue. An MQ local queue that holds messages in transit between queue managers. One transmission queue per target queue manager (or one per channel in some configurations). Chapter 19.

Trigger (MQ). A mechanism where the arrival of a message on a queue automatically starts a program or CICS transaction to process it. Trigger types: FIRST (trigger once when queue goes from empty to non-empty), EVERY (trigger for each message). Chapter 20.

TWS (Tivoli Workload Scheduler). IBM's enterprise job scheduling product for z/OS (formerly OPC). Manages batch job dependencies, resource allocation, and scheduling. Chapter 23.

U

UACC (Universal Access Authority). A RACF profile attribute that defines the default access level for users not explicitly permitted. Should always be NONE for production profiles. Chapter 28.

UDF (User-Defined Function). A function registered in DB2 and written in COBOL (or other languages) that can be called from SQL statements. Types: scalar (returns one value), table (returns a result set). Chapter 10.

UR (Uncommitted Read). A DB2 isolation level that reads data without acquiring locks, including uncommitted changes by other programs. Fastest but provides no consistency guarantees. Chapter 8.

V

Velocity goal. A WLM service class goal that measures the percentage of time a workload is not delayed by resource contention. Used for work that doesn't have a meaningful response time target. Chapter 5.

Virtual storage. The z/OS memory abstraction that gives each address space the illusion of a large, contiguous memory space, independent of physical memory. z/OS maps virtual addresses to physical storage and paging datasets. Chapters 1, 2.

Volume elasticity. The ratio of processing time increase to volume increase for a batch job. An elasticity of 0.9 means a 10% volume increase causes a 9% time increase. Chapter 23.

VSAM (Virtual Storage Access Method). The z/OS access method for keyed, sequential, and relative record datasets. Types: KSDS, ESDS, RRDS, linear. Chapters 4, 13.

W

WLM (Workload Manager). The z/OS component that manages system resources to meet service-level goals. Classifies work into service classes, sets dispatching priorities, and manages enclaves. Chapter 5.

WLM-managed address space. A z/OS address space whose lifecycle is managed by WLM. Used for DB2 stored procedures — WLM starts and stops instances based on demand. Chapter 10.

WITH HOLD (cursor). A DB2 cursor attribute that preserves cursor position across COMMIT points. Essential for batch programs that commit periodically within a cursor loop. Chapters 8, 12, 24.

X

XA (Extended Architecture) transaction. A distributed transaction that coordinates multiple resource managers (e.g., DB2 and MQ) under a single commit/rollback. In CICS, managed by the CICS recovery manager using two-phase commit. Chapter 18.

XRC (Extended Remote Copy). IBM's asynchronous disk replication technology. Provides near-zero data loss at any distance. Foundation for GDPS/XRC. Chapter 30.

Z

z/OS. IBM's flagship mainframe operating system. Successor to OS/390 and MVS. Provides address space isolation, hardware-enforced security, Parallel Sysplex clustering, and sub-second transaction processing for enterprise workloads. Chapter 1.

z/OS Connect. IBM's product for exposing z/OS assets (CICS programs, IMS transactions, DB2 stored procedures) as RESTful APIs with OpenAPI specifications. Chapters 21, 33.

zIIP (z Integrated Information Processor). A specialized processor on z/Architecture mainframes that executes eligible workloads (DB2, Java, XML, IPSec) at no additional MSU software cost. Chapter 29.

Zowe. An open-source framework for z/OS that provides a modern CLI, APIs, and a web-based desktop for interacting with z/OS resources. Part of the modern mainframe development toolkit. Chapter 36.