RESOURCES
GLOSSARY
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
=active log - Portion of the DB2 log to which new log records are written. The active log always contains the most recent log records. Older records are migrated to the archive log.
ACS (STK) - Automatic Cartridge System. Contains at least one LSM.
address space - The area of virtual storage available for a particular job.
address space connection - The connection between an allied address space and DB2. Each address space with a task connected to DB2 has a single address space connection, even though more than one task control block (TCB) might be involved.
allied address space - An address space connected to DB2 that is capable of requesting DB2 services.
allied thread - A thread from the local DB2 subsystem that may access data from a remote DB2.
application plan - Control structure produced by bind that is used by DB2 to process SQL statements.
archive log - Portion of the DB2 log that contains records migrated from the active log.
ATL - Automatic Tape Library.
attachment facility - Interface between DB2 and allied address spaces (e.g. TSO, IMS, CICS or batch). Allows application programs to access DB2.
authorization ID - An identifier that DB2 can validate for connection that has associated processing capabilities. It can relate to an individual, a group or a function. Installation conventions determine its meaning. Part of the correlation header.
availability - Proportion of time a system can be used for productive work
bind - The process that converts DB2 precompiler output to a usable control structure called a package or an application plan.
BMP - IMS Batch Message Processing.
buffer pool - Main storage area used to contain buffers for one or more table spaces or indexes.
cache - A high-speed buffer storage that contains frequently accessed instructions and data; it is used to reduce access time.
A design technique that makes a slower function effectively much faster. With CPU design, refers to high-speed memory in a CPU that contains current data and instructions from central storage. With DASD it is typically a solid-state memory containing currently active data from real DASD. The DASD component of virtual tape is often referred to as a cache.
capacity planning - The planning of computer resources to insure that workload service-level objectives can be met in the future. The computer resources contemplated in this definition are the CPU, real storage (central and expanded), the paging subsystem and the DASD I/O subsystem.
capacity planning horizon - The time period over which capacity is to be forecasted. Typical capacity planning horizons range from six to 24 months but can be as long as three or four years.
capacity threshold - The level of CPU utilization (defined in terms of the current CPU) that, if exceeded during the planning interval, will require the acquisition of additional CPU resources. Typical capacity thresholds range from 75% to 90%. The selection of the threshold is influenced by the relationship of the peak hour CPU utilization to the planning interval CPU utilization.
capture ratios - The ratio of measured (captured) CPU time to the total CPU time actually used. For CICS the measured workload time is typically referred to as user time.
Central Electronic Complex (CEC) - A CEC is a specialized device that performs all of the high speed computing associated with a computer. The IBM z990 CEC cage for example contains multiple central processing units.
central storage - Storage that is an integral part of the processor unit. Central storage includes both main storage and the hardware system area.
checkpoint - The act of recording status information to the DB2 log that would be needed for recovery if DB2 abended.
CICS™ (Customer Information
Control System) - An online transaction processing (OLTP) program from IBM
that, together with the COBOL programming language, has
formed over the past several decades the most common set
of tools for building customer transaction applications in the
world of large enterprise mainframe computing.
CICS attachment facility - DB2 component that provides linkage between DB2 and CICS.
claim - Indicate to DB2 that an object is being accessed. That object cannot be drained until a commit is reached.
commit - An operation that ends a unit of work. Locks are released, and database changes made by that unit of work can be referenced by other requests. A transaction can contain multiple commits.
connection name - An identifier supplied by the attachment facility that is associated with a specific address space connection. Part of the correlation header.
conversation - A dialog between two application processes on different DB2 subsystems. Might involve different MVS images.
correlation header - An identification section that is included with DB2 SMF 101 accounting records.
correlation ID -
An identifier of a specific thread. Part
of the correlation header.
Correlation Analysis - The capability to investigate possible relationships between performance metrics in your system.
Direct Access Storage Device (DASD) - A storage device in which the locations of each data record can be directly addressed. A device in which access time is effectively independent of location of the data.
DB2™ - A family of relational database management system (RDBMS) products from IBM that serve a number of different operating system platforms.
DASD Buffer (STK VSM) - The RAID DASD of a VTSS that contains current data. It is logically between the physical tape cartridges and the host channels. New logical tape volumes (VTV) are written from host channels to this DASD. Existing volumes are read from this DASD, potentially after being recalled from real tape (MVC).
DBMS - Database management system.
database - For DB2, refers to a collection of tables, or table spaces and index spaces.
database access thread - For DDF processing, refers to a thread accessing data locally on behalf of a remote subsystem.
data sharing - The ability for multiple DB2 subsystems to directly access and change a single set of data. Can involve one or more MVS images.
DBAT - Database access thread. A thread used to access DB2 data for a remote subsystem.
DBRM - Database request module. Contains information about a set of SQL statements. Used in the bind process.
DDF - Distributed database facility. Used by DB2 to communicate with another RDBMS.
deadlock - Irresolvable contention for a requested resource.
degree of parallelism - The number of concurrent operations used to process a specific query.
DL/I - Data Language/I. Data manipulation language for IMS databases.
Drain - Acquire a locked resource by quiescing other access to it.
drain lock - A lock that prevents a claim from occurring.
DRDA - Distributed relational database architecture.
entity - An item selected in the Instance Selector. Selected items are highlighted in the list box. An entity can be a buffer pool, DDF location or application group. An item selected in the I/O Selector. Selected items appear in the right-hand list box. An entity can be a device, a string, an LCU, a Storage Group, or a user defined group.
expanded storage - An extension of real storage available in IBM 3090 processors and above.
Gigabyte (GB) - A unit of measure for storage size. One gigabyte equals 1,073,741,824 bytes or 1,024 MB.
group - A user-defined set of transaction groups identified by a unique one- to eight-character name.
group set - One or more user-defined groups saved or restored by a specific name.
grouper - An interface that allows a user to group and select workloads for viewing and analysis. A user interface that allows you to analyze device specific metrics and to define groups of devices.
history charts - Graphical representations of performance data that show metrics as variables and plot as observations the planning interval averages for each selected period of data.
HSC - Host Software Component. The host software components of an STK Nearline implementation.
import - The process of identifying a file as belonging to a system and having a period name within a period type. For PerfMan for DB2, this is accomplished with PMDLOAD.
I/O rate - The average number of I/Os occurring per second.
I/O subsystem - The configuration of channel paths, controllers (both cached and non-cached) and DASD devices in an MVS environment that handles the transfer of data between central storage and the peripheral equipment.
IMS™ (Information Management System) - Database and transaction management software that was first introduced by IBM in 1968. IMS is one of two major legacy database and transaction management subsystems from IBM that run on mainframe MVS (now z/OS) systems.
IRLM - Internal resource lock manager. An OS/390 subsystem that DB2 uses to control communication and locking.
Kilobyte (KB) -
A unit of measure for storage size, One Kilobyte equals 1024
bytes.
Latent Demand - The demand for system resources that goes unsatisfied for various reasons. For example, users may tend not to use the system in the afternoon because their experience has been that response time is very poor then. Or slow response time may inhibit a data entry operator from entering transactions at his normal rate. Changes to the system may allow it to satisfy this kind of latent demand. If utilization on an upgraded system seems higher than what was expected, it may be due to latent demand being satisfied.
LCU (STK) - Library Control Unit. Portion of an LSM that controls robotics.
Logically Partitioned (LPAR) mode - A CPC power-on reset mode that enables use of the processor feature and allows an operator to allocate CPC hardware resources (including central processors, central storage, expanded storage, and channel paths) among logical partitions.
LMU - Library Management Unit. The part of an STK ACS that communicates with the host and manages up to 16 LSMs.
LRU - Least recently used. Default algorithm often used to manage use of real resources by virtual things. With computers, it refers to the management of virtual pages that reside in real memory. With virtual tape, it controls which virtual volumes reside on VT DASD. When real space is required, virtual things that have gone the longest period of time without being referenced are the first deleted. VT performance is enhanced when enough DASD is available to let virtual volumes stay resident long enough to be reused. It is also enhanced if there are techniques to remove volumes that are not likely to be reused.
LSM - Library Storage Module. Storage area for tape cartridges and the robot used to access them.
Linear Tape-Open (LTO) - An open-format tape storage technology developed by Hewlett-Packard (HP), IBM and Certance. The term "open-format" means that users have access to multiple sources of storage media products that will be compatible.
main storage - Program-addressable storage from which instructions and other data can be loaded directly into registers for subsequent execution or processing.
Megabyte (MB) - A measure of computer processor storage and real and virtual memory. One megabyte is 2 to the 20th power bytes, or 1,048,576 bytes.
MIPS ( Million Instructions Per Second) - A method for measuring the Capacity or throughput capability of a mainframe system.
migrate - The virtual tape function of copying new or changed virtual tape volumes from VT DASD to real tape (MVC). Depending on implementation this might take place as soon as the new volume is closed, or it might be scheduled later. After completion, the VTV might also remain on DASD for potential reuse.
Modeling - The ability to take a given workload and simulate this workload running on a Configuration, in order to analyze what the impact of running it on this configuration will be.
MSU (
Millions of Service Units) - A methodology for measuring utilization of
a mainframe System. Similar to MIPS (Millions Instructions Per Second).
MVC (Multi Volume Cartridge) - High capacity tape cartridges that contain images of logical volumes. Typically many logical volumes will be stacked on a single cartridge.
Nearline (STK) - Tape automation. The combination of hardware and software that manages tape libraries and robotics.
Network-attached storage (NAS) - Hard disk storage that is set up with its own network address rather than being directly attached to a computer.
observations - Individual values of a variable plotted along the X-axis. Typically each variable will be shown with the same number of observations.
Operating System (O/S) - Software that controls the execution of programs and that may provide services such as resource allocation, scheduling, input/output control, and data management.
package - A set of statically bound SQL statements that is ready for processing. Also referred to as an application package.
paging - The movement of individual pages (4,096 bytes each) between real storage and auxiliary storage (the page data sets) or expanded storage because there is insufficient real memory to hold all the pages needed. The CPU can only execute instructions and operate on data in real memory. Without enough real memory, the operating system writes old pages to disk and reads new pages from disk so that needed data and instructions are in memory.
paging subsystem - The paging set of auxiliary storage devices and expanded storage used for paging and swap paging.
performance management - The management of computer resources to meet workload service-level objectives. The computer resources contemplated by this definition are the CPU, real storage (central and expanded), the paging subsystem and the DASD I/O subsystem.
period - A unique segment of time for which data has been accumulated into a file.
period charts - Graphical representations of performance data that show metrics as variables and plot as observations the metrics' value for each interval of the current period.
period type - A logical grouping of periods that have the same summarization characteristics.
planning interval - That time period during the day for which it is most important that adequate capacity exist to meet service-level objectives. It may be as short as the peak hour of the day or it may be as long as a 12-hour shift. The planning interval may also be non-contiguous. This subset of time within the 24 hours of a day is used to compute an average number.
queue time - The amount of time in milliseconds that a server delays a request due to an outstanding request to a common resource.
real storage - The main storage in a virtual storage system. Physically, real storage and central storage are identical. Conceptually however, real storage represents only part of the range of addresses available to the user of a virtual storage system. Real storage is the single most critical resource to any operating system.
relative I/O content - This processor-independent value represents the "I/O-ness" of your system. It can be considered the I/O rate of your system per processor percentage used. The higher the RIOC, the more I/O-intensive your system. The lower the RIOC, the more CPU-intensive.
response time - The average time required to complete a function or transaction. Typically refers to the time within DB2 for an application request.
response time profile (R/T profile) - Represents the component parts of response time, including CPU and various wait conditions.
RMIP S (Relative RMIPS) - Enables comparison of processing capability between various processors.
recall - The VT function of copying a logical tape volume from real tape (MVC) to DASD.
reclaim - The VT function of coping one MVC to another to eliminate obsolete data. When a new version of a virtual tape dataset is written by host software, the previous MVC becomes obsolete.
Remote Monitor Facility (RMF) - System activity records or logs that are captured and recorded by the os/390 and z/Series mainframes.
RTD - Real Tape Drive. Physical tape drives, typically of high capacity, used to transfer logical tape volumes between VTS or VTSS DASD and tape cartridges (MVC). These devices are not seen by host applications.
service-level objectives - Criteria by which to judge whether a system is meeting its performance objectives, usually defined in terms of average response time for interactive and batch workloads, and average unit response time for all other workloads.
With tape subsystems, logical measurements include mount time, through-put, and I/O rates.
service time - The amount of server time required to satisfy a request. For example, CPU service time for a transaction is the amount of CPU time the transaction uses.
SLF - Server log file. The host file produced by DB2SERVE (or DB2WIPSV) for subsequent processing by PerfMan Server.
The host file produced by programs ISMVTS (for IBM VTS and libraries) and ISMVSM (for STK Nearline and VSM).
Storage Area Network (SAN) - A SAN is a high-speed special-purpose network (or subnetwork) that interconnects different kinds of data storage devices with associated data servers on behalf of a larger network of users..
thread - Connection between an application and DB2.
transaction - For most interactive work (e.g. IMS, CICS, DDF) each logical transaction results in one SMF 101 record. PerfMan for DB2 considers each SMF 101 record to be a transaction for tread groups.
Tape Volume Cache (TVC) - The RAID DASD in the VTS that is where current data resides. It is logically between the physical tape cartridges and the host channels. New logical tape volumes (VTV) are written from host channels to this DASD. Existing volumes are read from this DASD, potentially after being recalled from real tape (MVC).
uncaptured time - In general, the portion of total time that cannot be specifically associated with a function. Typically total elapsed can be measured. Often individual components of the total can also be measured. Uncaptured (often "Other" on PerfMan charts) is the difference between total elapsed and sum of components.
In general, the sum of the CPU time reported (captured) by a measurement facility is less than the total CPU time actually used. The difference between the total CPU time actually used and captured time is called uncaptured time.
unit response time - The total elapsed time of the address space in the performance group divided by the unit transaction count. Used to define a standard unit of work across all selected groups to enable direct comparison.
unit Transaction - That unit of work that executes 500,000 instructions. It is defined at the performance group level, and used to define a standard unit of work across all performance groups to enable direct comparison with other performance groups.
variables - Measurements/metrics illustrated on a chart. Typically, a variable has a unique occurrence in the chart legend.
virtual storage - The storage space that may be regarded as addressable main storage by the user of a computer system in which virtual addresses are mapped into real addresses. The size of virtual storage is limited by the addressing scheme of the computer system and by the amount of auxiliary storage available - not by the actual number of main storage locations.
virtual tape - Any of several techniques for simulating tape volumes and drives with a combination of hardware, microcode, and software. Two specific examples are IBM VTS and StorageTek VSM. To host applications, they look like groups of 3490E tape drives. They can do an excellent of solving traditional tape problems - few records on a tape, long mounted times with little data transferred, wanting to mount more tapes than available drives.
VSM - Virtual Storage Manager. StorageTek implementation of virtual tape. Consists of hardware (VTSS) and software (VTCS), added to a Nearline environment.
VT - Virtual Tape.
VTCS - Virtual Tape Control System. The VSM HSC component.
VTD - Virtual Tape Drive.
VTS - Virtual Tape Server. Specifically refers to the IBM virtual tape hardware. Sometimes used generically to refer to virtual tape systems.
VTSS (STK) - Virtual Tape Storage Subsystem. The hard component that contains DASD buffer and microcode to appear to the host as real tape.
VTV - Virtual Tape Volume.