MVS

Session 1
Classification of computers…
 Computers are divided into three classes:
◦ Micro-computers,
◦ Mini-computers and
◦ Mainframe.

 The division is loosely based on the SIZE of the

computer, which depends on:
◦ Its hardware architecture,
◦ The complexity of its system software and
◦ The nature of its application programs.

© Strictly confidential 2
…Classification of computers…
Micro-computer

Hard disk
1.2 to 20 GB
Terminal Processor
32 MB to 256 MB
Monitor
Diskette
+
1.44 MB
Keyboard

Printer

© Strictly confidential 3
…Classification of computers…
Mini-computer

Terminal Disk drive

10 to 20 GB

Terminal Processor Disk drive

128 MB to 512 MB 10 to 20 GB
Terminal
Disk drive
10 to 20 GB
Terminal
Printer Printer
Disk drive
Terminal
10 to 20 GB

© Strictly confidential 4
…Classification of computers…
Mainframe
Disk drives
Tape drives ( 10 GB to 40 GB each )
Terminals

Terminals
Processor
Terminals
512 MB to 2 GB
Terminals

Terminals
Printer Printer Printer

© Strictly confidential 5
Basic components of a computer

I/O DEVICES
PROCESSOR

Terminals

CPU
Printers
Device
Controllers
Main Disk drives
Storage

Tape drives

© Strictly confidential 6
History of IBM Mainframes…
 Scientific Computers
◦ Floating point arithmetic
◦ Examples: IBM 704, 709, 7090 and 7040 - 44

 Decimal Computers
◦ Decimal arithmetic
◦ Examples: IBM 7070 and 7080

 Character Computers
◦ Character oriented
◦ Examples: IBM 1401, 1410 and 7010

© Strictly confidential 7
…History of IBM Mainframes…
 System 360
◦ Released in 1964
◦ Unified machine architectures
◦ Uni-processing machines
◦ Programs are upward compatible

 System 370
◦ Released in 1972
◦ Facilitates DAT
◦ Multiprocessing machines
◦ Extended Real Addressing

© Strictly confidential 8
…History of IBM Mainframes
 System 370 XA
◦ Released in 1981
◦ Natural extension of System 370
◦ Address space expanded to 2 GB from 16 MB
◦ Addition of substantial new hardware to manage
multiprocessor

 System 370 ESA

◦ Released in 1988
◦ Addition of new hardware to support data spaces

 System 390
◦ Released in 1990
◦ Introduced ESCON architecture

© Strictly confidential 9
Evolution of MVS…
 Primary Control Program – PCP
 Multiprogramming with Fixed number of Tasks – MFT
 Multiprogramming with Variable number of Tasks – MVT
 Operating System / Virtual Storage-1 – OS / VS 1
 Operating System / Virtual Storage-1 – OS / VS 2 or
Single Virtual Storage – SVS
 Multiple Virtual Storage – MVS
 Multiple Virtual Storage / Extended Architecture – MVS /
XA
 Multiple Virtual Storage / Enterprise System
Architecture –
MVS / ESA

© Strictly confidential 10
…Evolution of MVS

Ave. no. of
Operati jobs Virtual
Hardwar Storage
ng running Storage
e Types
System concurrent Limits
ly
PCP S / 360 1 Main Not available
MFT and S / 360 10 to 20 Main Not available
MVT
OS / 1 S / 370 30 to 50 Main and 16 MB
and SVS Auxiliary
MVS S / 370 100s Main and 16 MB
Auxiliary
MVS / XA S / 370 1000s Main, 2 GB
XA Expanded and
Auxiliary
MVS / S / 370 1000s Main, 2 GB + Data
ESA ESA Expanded and and Hiper
© Strictly confidential
Auxiliary spaces 11
Features of MVS…
 Virtual Storage
 Multiprogramming
 Spooling
 Batch Processing
 Time Sharing (1971)

VIRTUAL STORAGE
 Is a facility provided to make efficient use of the main
storage.
 Is a technique that lets a processor simulate a large
amount of main storage from a smaller installed main
storage.
 Implementation: Uses disk storage as an extension of
main storage.
 Key: At any point in time, only one program and its data
is required in the main storage.
© Strictly confidential 12
…Features of MVS…
MULTIPROGRAMMING
 Is a facility provided to make efficient use of the CPU.
 Is a technique that lets more than one program run at
the same time.
 Key: I/O devices are much slower than Processor.

SPOOLING
 Is a facility provided to make efficient use of the output
devices.
 Is a technique that intercepts the printer output and
re-directs to a disk file.
 Each program’s output is stored separately.
 Facilitates effective sharing of output devices and
provides better system throughput.

© Strictly confidential 13
…Features of MVS
BATCH PROCESSING
 Is a technique by which any task is accomplished in units
called jobs.
 The necessary information for batch processing is provided
through JCL.
 Is managed by the system’s job scheduler which makes
decision about the order in which the jobs should be
executed based on the job priority and class.
 Is a non-interactive and offline mode of data processing.

TIME SHARING
 Is a technique by which the interactive processing is
facilitated.
 Each user accessing the system through a terminal device
gets a time slice repeatedly.
 At the background the processing is still done in the batch
mode.
© Strictly confidential 14
Storage
DASD
 Is the official IBM term for a disk drive.
 Provides rapid and direct access to large quantities of
data.
 Read and write data on a disk pack (sometimes called a
volume), which is a collection of metal platters coated
with a metal oxide.

© Strictly confidential 15
Address space
 Is the complete range of addresses that can be
accessed by a processor for a particular program.
 Depends on the number of bits used for addressing:
◦ System 370 is a 24 bit processor and hence can
address only 24 MB.
◦ System 370 / XA, 370 / ESA and ES 9000 are 31 bit
processors and hence have an addressability of 2 GB.

 From MVS/ESA, data space and hiper space are

provided to augment address space.

© Strictly confidential 16
Multiple Virtual Storage…
 MVS simulates several address spaces, each
independent of the other that represents a separate
user, job or a subsystem.
 Is supported by the following mechanisms:
◦ Paging.
◦ Swapping.

© Strictly confidential 17
…Multiple Virtual Storage
Address space 1 Address space 2 Address space 3

Page
Frame

DASD Volume

Page Slots Page

© Strictly confidential 18
Paging
 Pages are 4 KB blocks in virtual storage.
 Page frames are 4 KB blocks in real storage.
 Page table maps Pages to Page frames.
 Page fault occurs when a program refers to data not
available in real storage.
 Page-in when MVS loads a new page into a page frame.
 Page-out when data in a page-frame is written back to
DASD.

© Strictly confidential 19
Swapping
 Refers to transferring of an entire address space in and
out of virtual storage.

© Strictly confidential 20
Expanded storage
 Large buffer between real storage and page data sets
(Virtual Storage).
 Improves the efficiency of virtual storage operations.

© Strictly confidential 21
Program execution modes
 Real Mode
◦ Refers to execution of programs that can not be
subject to Paging and Swapping.
◦ Examples: MVS routines that handles Paging and
Swapping and other OS routines.

 Virtual Mode
◦ Refers to execution of programs that can be subject
to Paging and Swapping.
◦ Examples: Application programs.

© Strictly confidential 22
Address space, Data space, Hiper space
Address space Data space
Every application program can An application program can
have only one address space. have multiple data spaces (up
to 7999 per application
program).
Can store both data and Can store only data. Even if
programs. programs are stored they are
regarded as data.
System area and Common System area and Common
areas are mapped onto address areas are not mapped onto
space. Data space data space.
Hiper space
Backed by Main, Expanded and Backed by Expanded and
Auxiliary storages. Auxiliary storages.
Byte addressable. Page addressable.
Data is sharable. Data is non-sharable.
Data is accessible to Data has to brought to data
application programs. space or address space for
accessing.

© Strictly confidential 23
Data set
 A data set is an IBM term for a file.
◦ Within a data set, data is organised into smaller units
called records, which can be processed individually
by application programs.
◦ MVS keeps track of data through labels, catalogues
and different data set organizations

© Strictly confidential 24
Data set organisations
 Non-VSAM
◦ Physical Sequential (PS)
◦ Partitioned Data Set (PDS)
◦ Index Sequential Access Method (ISAM)
◦ Direct

 VSAM
◦ Entry Sequential Data Set (ESDS)
◦ Key Sequence Data Set (KSDS)
◦ Relative Record Data Set (RRDS)
◦ Linear

© Strictly confidential 25
Non-VSAM data set organisations…
Physical Sequential (PS)
 Records are stored one after the other in a sequence.
 Facilitates only sequential access to the records.
 Can reside on any kind of storage.

Partitioned Data Set (PDS)

 Is divided into one or more members, each of which can
be processed independently.
 Each member name is stored in a directory.
 Is like a pseudo-directory.
 Also called a library.
 Widely used to store related programs and data.

© Strictly confidential 26
…Non-VSAM data set organisations
Index Sequential (ISAM)
 Essentially comprises of two components:
◦ Index component.
◦ Data component.
 Provides both sequential and random access to records.
 Random access is facilitated through the index
component.

Direct
 Facilitates random access based on record’s disk
location.
 Is rarely used due to the complexity involved in
computing the record’s disk location.

© Strictly confidential 27
VSAM data set organisations
 ESDS: Similar to PS.
 KSDS: Similar to ISAM.
 RRDS: Similar to Direct.
 Linear: Line sequential.

© Strictly confidential 28
Data access methods
 Data Access methods are MVS routines that facilitate
the file and database operations such as OPEN, CLOSE,
READ, WRITE, RE-WRITE, DELETE, etc.
 There are two types of data access methods:
◦ Data set Access methods: For file operations.
◦ Database Access methods: For database operations.

© Strictly confidential 29
Record formats in MVS
 Fixed Unblocked
 Fixed Blocked
 Variable Unblocked
 Variable Blocked
 Undefined

© Strictly confidential 30
Job management
 A job is a collection of one or more related programs.
Each program in a job is called a job step.
 The specifications necessary to process a job are
provided though a JCL.
 Job management is the task of a MVS subsystem called
JES. Its responsibilities are:
◦ Entering the job into the system.
◦ Decide when the job is to be processed.
◦ Pass the job to MVS for processing.
◦ Delivering the job’s output to correct destination.

© Strictly confidential 31
Job processing…
 Job submission:
◦ Create a JCL using the ISPF.
◦ Submit the JCL using “submit” or “sub” command.
◦ JES copies the job stream into the JES spool.

© Strictly confidential 32
…Job processing…
 Job selection:
◦ Is based on job class and job priority.
◦ Is assigned to an initiator that is responsible for
running a batch job.
◦ An initiator is a program that runs in the system
region of the address space.
◦ Each initiator can execute one program at a time and
the number initiators can be varied dynamically.
◦ Each initiator has one or more job classes associated
with it.
◦ Within an initiator with specified classes, job selection
is made on the job priority (0 to 15).
Continued…

© Strictly confidential 33
…Job processing…
 Job selection (continuation):
Job
Characteristics
class
A Will execute within 15 minutes of submission
B Will execute within 30 minutes of submission
C Will execute within 1 hour of submission
D Will execute overnight
H Will be held until released by operator
L Within 15 minutes of submission, each step limited to
1 min execution time
T Requires tape processing
Initiato Eligible job
rs classes
1 A
2, 3 B, C, D, H, L, T
4, 5 B, C
5 B, C
© Strictly confidential 34
…Job processing…
 Job execution:
◦ Initiator picks up a job for execution and invokes the
interpreter.
◦ The interpreter checks the JCL for syntax errors and
creates a series of control blocks required for the job.
◦ The initiator does the following:
 Invokes the allocation routines.
 Builds the user address space and executes.
 Invokes the de-allocation routines.

© Strictly confidential 35
…Job processing
 Processing the job output:
◦ Like jobs, the output is also assigned to different
output classes:
 Class A: Standard printer output.
 Class Z: Held output.
◦ Single job may process output using more than one
output class:
 Class A: System messages, program output.
 Class D: Job’s spooled output.
◦ Specifications for output like number of copies, etc.

 Job purging:
◦ JES spool space held by the job is released.
◦ JES control blocks held by the job are deleted.

© Strictly confidential 36
System data sets
 SYS1.NUCLEUS - MVS nucleus.
 SYS1.PARMLIB - Options for MVS initialization.
 SYS1.LINKLIB & SYS1.LPALIB - MVS executables.
 SYS1.MACLIB - MVS macro instructions.
 SYS1.PROCLIB - JCL procedures.
 SYS1.CMDLIB - TSO commands.
 SYS1.LOGREC - Info about hardware problems.
 SYS1.DUMP01 - Storage dumps on MVS component
failure.
 Page data sets:
◦ Hold individual pages of virtual store for paging
support.
◦ 8 to 10 data sets of 250 to 300 MB.
 Swap data sets:
◦ Hold some pages when entire address space is
swapped out. © Strictly confidential 37
MVS subsystem
 A subsystem is a software product that operates in a
separate address space under the control of MVS.
 Within its address space a subsystem may provide
facilities that duplicates the operating system’s
facilities.
 Some important MVS subsystems:
◦ TSO and ISPF.
◦ CICS and IMS/DC.
◦ DB2, IDMS and IMS/DB.
◦ SMS and JES.
◦ VTAM.
◦ SMF.

© Strictly confidential 38
Other components and facilities
 Language translators, Linkage editors and Loaders.
 General Purpose utility programs.
 REXX and CLIST.
 RACF.
 SDSF.

© Strictly confidential 39