This primer on z/OS performance terminology is provided by the experts at PerfTechPro. Like any technical field, this one is filled with acronyms and over-used terms. We provide this primer to aid those who may be new to the profession.
| Term | Explanation |
|---|---|
| Address Space | An Address Space in a z/OS system is a logical construct in which programs execute in a z/OS system. An address space gives the program a complete and isolated view of the virtual memory in which the program is executing. |
| Address Space Status | Address Spaces that are active in a z/OS system have a particular status at any particular time. They may have status: In and Ready, In and Wait, Out and Ready, Out and Wait, Logical Ready and Logical Wait. |
| Address Space Type | Address Spaces that are active in a z/OS system are of a particular type. They may have type: STC, Batch, TSO, OMVS and ASCH. The type of an address space generally describes the type of work that runs in the address space. The most common type of address space is STC under which z/OS operating system functions run. User work generally runs in Batch and TSO address spaces. |
| BATCH | Batch refers to one of the basic types of work (along with TSO and Started Tasks) that runs in a z/OS operating system. A batch job is initiated by the job entry subsystem (JES2 or JES3) in use by the particular z/OS operating system. |
| CEC or CPC | A Central Electronic Complex (CEC) is a term often used to refer to an IBM mainframe computer. Another term used less frequently to refer to an IBM mainframe computer is Central Processor Complex (CPC). |
| Central Storage | Central Storage is the physical memory in IBM mainframe computers used to store programs and data. |
| CHANNEL | A channel (or the channel subsystem consisting of multiple channels) in an IBM mainframe computer is hardware that initiates and manages IO operations independent of program execution on the mainframe itself. The purpose of the channel subsystem is to allow inherently slower (than instruction execution on mainframe processors) IO operations to proceed concurrently with mainframe processor instruction execution. |
| CICS | CICS (Customer Information Control System) is a family of IBM products designed as middleware to support rapid, high-volume online transaction processing. |
| Coupling Facility | A coupling facility enables software on different systems in the Parallel Sysplex to share data with the assurance that the data will not be corrupted and will be consistent among all sharing users. To share data, systems must have connectivity to the coupling facility through coupling-facility links. Coupling Facility (CF) provides locking, caching, and list services between coupling-capable z/OS processors. Coupling-facility links connect a coupling facility to the coupling-capable processors. The coupling-facility control code (CFCC) provides the coupling-facility functions. The CFCC can run in a logical partition (LPAR) under PR/SM. |
| DASD | A direct-access storage device (DASD) is a storage device in which each physical record has a discrete location and a unique address. The term was coined by IBM to describe devices that allowed random access to data. |
| Expanded Storage | Expanded Storage was a form of physical memory in IBM mainframe computers. It is no longer used in IBM processors. |
| GPP | A General Purpose Processor (GPP) is a type of central processor in an IBM mainframe computer that can execute instructions for any type of work running in an LPAR. |
| In and Ready Address Spaces | This number represents the average number of address spaces that are either dispatched on logical processors (LPs) or waiting to to be dispatched on LPs. The average number of In and Ready Address Spaces is a key measure of contention for the GPP resource in a z/OS system. |
| JES2 | z/OS uses a job entry subsystem (JES) to receive jobs from a job queue, schedule jobs for processing, and control job output processing. JES2 (Job Entry Subsystem 2) is a job entry subsystem used by z/OS that is descended from HASP (Houston Automatic Spooling Priority), which was a computer program that provides supplementary job management, data management, and task management functions such as: scheduling, control of job flow, and spooling. HASP persists within JES2 as the prefix string for most module names and for all messages sent by JES2 to the operator. |
| LP (Logical Processor) | A Logical Processor (LP) is the logical representation of a central processor (CP) in a logical partition (LPAR). The operating system running in an LPAR dispatches work to run on a logical processor.. PR/SM then associates a physical processor with the logical processor to actually execute the instructions required by the work dispatched on the logical processor. Each LPAR has a certain number of logical processors assigned to it. In the case of an LPAR running the z/OS operating system, these logical processors may be general purpose LPs or zIIP LPs. |
| LPAR (Logical Partition) | A Logical Partition (LPAR) is a subset of a computer’s hardware resources, virtualized as a separate computer. In effect, an IBM mainframe computer can be partitioned into multiple logical partitions running under the control of PR/SM, each hosting a separate instance of an operating system like z/OS, z/VM or zLinux. |
| LSPR | LSPR stands for ‘Large Systems Performance Reference’. It is the banner under which IBM publishes the results of its mainframe benchmarks.. The results of these benchmarks are published in terms of Internal Throughput Ratios (ITRs). ITRs are relative capacity measures for IBM mainframes. For example, a processor with an ITR of 2.0 would be considered to have twice the capacity of a processor with an ITR of 1.0. |
| MIPS | MIPS is a relative measure of the capacity of an IBM mainframe computer. Each mainframe will have a MIPS rating for its GPPs and a MIPS rating for its zIIPs. Originally, MIPS was the abbreviation for ‘Millions of Instructions Per Second’. That meaning of the term MIPS has been obsolete for some time. The MIPS rating (either for its GPPs or zIIPs) is based on benchmarks that IBM runs and publishes under the name ‘Large Systems Performance Reference’ or LSPR. |
| MSUS | MSUS is also a relative measure of the capacity of an IBM mainframe computer, but unlike MIPS, MSUS refers only to the capacity of the GPPs in the particular mainframe. The measure of the consumption of MSUS by workloads running under z/OS in an LPAR is used primarily for software pricing. |
| MVS | MVS (Multiple Virtual System) was the name of the second incarnation of a virtual memory operating system for IBM mainframe computers. The first was SVS (Single Virtual System). The acronym MVS is still often used to refer to the z/OS operating system. |
| Page Datasets | Page datasets are z/OS datasets that are used to store 4K pages of virtual storage that do not currently reside in 4K frames of central storage. When a page fault occurs, z/OS must retrieve the required 4k page from the 4K slot in a page dataset and move it into a 4K frame in central storage. |
| Paging | During program execution, only those 4K pages of virtual storage that are required reside in 4K frames in central storage. The pages remain in central storage until no longer needed, or until another page is required by the same application or a higher-priority application and no empty central storage is available. To select pages to be written to auxiliary storage, z/OS follows a “Least Used” algorithm. That is, z/OS assumes that a page that has not been used for some time will probably not be used in the near future. When a program references a page of virtual storage that is not backed by a frame in central storage, a page fault occurs. This requires z/OS to retrieve the page from auxiliary storage and bring it into central storage. |
| PR/SM | Processor Resource/System Manager (PR/SM) is a hypervisor that allows multiple logical partitions to share the physical resources of an IBM mainframe computer such as CPUs, I/O channels and LAN interfaces. PR/SM is integrated with all IBM System z mainframe computers. |
| RC (Report Class) | A Report Class is a category of work defined by the user in the Service Policy for the z/OS system running in a particular LPAR. A Report Class is a grouping of work that the user defines for reporting purposes only. For example, a report class might be defined in such a way as to report on the resource consumption of a particular collection of CICS user ids. |
| RMF | Resource Measurement Facility (RMF) is a measurement tool from IBM that monitors z/OS system activity to collect performance and capacity planning information. RMF Monitor I writes information on a user defined interval basis, typically every 15 minutes, to the SMF dataset. Although these records are technically SMF records, they are usually referred to as RMF records. |
| SC (Service Class) | A Service Class is a category of work defined by the user in the Service Policy for the z/OS system running in a particular LPAR. Service Classes may have multiple periods defined through which work running in a particular Service Class transitions during its execution. For example, a Service Class that runs TSO work typically has 3 or 4 periods defined. The work from a TSO user starts in period 1 and subsequently may transition to periods 2 or 3 or 4 depending on how long the work takes to finish. Importantly, the user defines service level objectives (SLAs) at the Service Class Period level. |
| SMF | IBM System Management Facility (SMF) is a component of IBM’s z/OS operating system for mainframe computers that provides a standardized method for writing records of system activity to a variable length sequential dataset. SMF provides records that fully describe all of the activities occurring under the z/OS operating system on an IBM mainframe. The records written by SMF to the SMF dataset are identified by record type. For example, the SMF type 70 record is written by RMF and contains configuration and performance information about LPARs, Logical Partitions, Address Spaces and more. |
| STC (Started Task) | The term ‘Started Task’ refers to a particular type of address space running under the control of the z/OS operating system. The basic units of work in a z/OS operating system are Started Tasks, Batch Jobs and TSO Users. Batch jobs are scheduled by a job entry subsystem (JES2 or JES3) and are scheduled to run based on the resources they require and their availability, or based on controls that you place on the batch system. A started task is initiated by JCL that is run immediately as the result of a START command. Started tasks are generally used for system functions and critical applications. An advantage offered by started tasks are control over where and when the JCL is run. For example, you could have the JCL started at each IPL of the system. Controlling where and when a batch job runs is more complex than using a started task. |
| TSO | Time Sharing Option (TSO) allows users to create an interactive session with a z/OS system. TSO provides a single-user logon capability and a basic command prompt interface to z/OS. Most users work with TSO through its menu-driven interface, Interactive System Productivity Facility (ISPF). ISPF provides a collection of menus and panels that offer a wide range of functions to assist users in working with data files on the z/OS system. |
| Virtual Storage | Virtual Storage in z/OS means that each program running under control of z/OS can assume it has access to all of the storage defined by the z/OS architecture’s addressing scheme. This ability to use a large number of storage locations is important because a program may be long and complex, and both the program’s code and the data it requires must be in central storage for the processor to access them. z/OS supports 64-bit long addresses, which allows a program to address up to 16 exabytes of storage locations. z/OS provides the functionality to have storage actually addressed by a running program in central storage when it is required. |
| Workloads | In setting up an instance of a z/OS operating system to run in an LPAR, the user defines a Service Policy. In that Service Policy, the user defines, among other things categories of work in terms of Service Classes, Report Classes and Workloads. In particular, a Workload consists of the work that runs in a collection of Service Classes. |
| zIIP | An Integrated Information Processor (zIIP) is a specialized type of central processor in an IBM mainframe computer designed to run only specific types of work, typically database work. The work that can run on zIIPs is referred to as zIIP eligible work. A zIIP processor is less expensive than a General Purpose Processor so the more work an installation can run on zIIP processors, the less expensive is the cost of running that work. |
IBM, the IBM logo, ibm.com, IBM Z, IBM z Systems, Parallel Sysplex, z Systems, z/OS, z/VM, zEnterprise are trademarks of International Business Machines Corporation, registered in many jurisdictions worldwide. A current list of IBM trademarks is available on the Web at “Copyright and trademark information” at: www.ibm.com/legal/copytrade.shtml