Operating

Systems:
Internals Chapter 2
and Design
Principles Operating System
Overview
Seventh Edition
By William Stallings
 Operating Systems:
Internals and Design Principles
Operating systems are those programs that interface the machine with
the applications programs. The main function of these systems is to
dynamically allocate the shared system resources to the executing
programs. As such, research in this area is clearly concerned with
the management and scheduling of memory, processes, and other
devices. But the interface with adjacent levels continues to shift with
time. Functions that were originally part of the operating system have
migrated to the hardware. On the other side, programmed functions
extraneous to the problems being solved by the application programs
are included in the operating system.

—WHAT CAN BE AUTOMATED?: THE COMPUTER SCIENCE AND

ENGINEERING RESEARCH STUDY,
MIT Press, 1980
 Operating System
 An interface between applications and computer
A program that controls the execution of
application programs and the allocation of
system resources
Main objectives of an OS:

• Convenience
• Efficiency
• Ability to evolve
Convenience:The OS as a
User/Computer Interface
 The OS provides abstractions of the computer
hardware, making it more convenient for
applications to use the computer’s capabilities
 It does this through a set of interfaces and
services.
 Key Interfaces

 Instructionset architecture (ISA)

 Application binary interface (ABI) supports
portability of applications in binary forms across different system
platforms and environments

 Application programming interface

(API) applications can interface to OS through system calls
 Program development utilities – not strictly OS
 Program execution
 Access I/O devices
 Controlled access to files
 System access
 Error detection and response
 Accounting
 Efficiency:The Operating System
As a Resource Manager

Acomputer is a set of resources for

moving, storing, & processing data
 TheOS is responsible for managing
these resources
 TheOS exercises its control
through software
 Functionsin the same way as ordinary
computer software
 Program, or suite of programs, executed
by the processor
 Frequently relinquishes control and must
be able to regain control to decide on the
next thing the processor should do.
Operating
System
as
Resource
Manager
Evolution of Operating
Systems
 A major OS will evolve over time for a
number of reasons:
Hardware upgrades

New types of hardware

New services

Fixes
 Evolution of
Operating Systems
 Stages include:

Time
Sharing
Multiprogrammed Systems
Batch Systems
Simple Batch
Systems
Serial
Processing
 Serial Processing
Earliest Computers: Problems:
 Scheduling:
 No operating system
 most installations used a
 programmers interacted
hardcopy sign-up sheet to
directly with the computer
reserve computer time
hardware
 time allocations could
 Computers ran from a console run short or long,
with display lights, toggle resulting in wasted
switches, some form of input computer time
device, and a printer
 Setup time
 Users had access to the
computer in “series”  a considerable amount of
time was spent just on setting
up the program to run
 Simple Batch Systems

 Early computers were very expensive

 important to maximize processor utilization

 Monitor (primitive operating system)

 user no longer has direct access to processor
 job is submitted to computer operator who batches them
together and places them on an input device
 Monitor controls the sequence
of events
 Resident Monitor is software
always in memory
 Monitor reads in job and gives
control
 Job returns control to monitor
 Processor executes instruction from the memory
containing the monitor
 Executes the instructions in the user program until it
encounters an ending or error condition
 “control is passed to a job” means processor is fetching and
executing instructions in a user program
 “control is returned to the monitor” means that the processor
is fetching and executing instructions from the monitor
program
Job Control Language
(JCL)
Special type of programming
language used to provide
instructions to the monitor

what compiler to use

what data to use

Memory protection for monitor

• while the user program is executing, it must not alter the memory area
containing the monitor

Timer

• prevents a job from monopolizing the system

Privileged instructions

• can only be executed by the monitor

Interrupts

• gives OS more flexibility in controlling user programs

 Modes of Operation

User Mode Kernel Mode

• user program executes in • monitor executes in kernel
user mode mode
• certain areas of memory are • privileged instructions may
protected from user access be executed
• certain instructions may not • protected areas of memory
be executed may be accessed
 Simple Batch System
Overhead
 Processor time alternates between execution of user
programs and execution of the monitor
 Sacrifices:
 some main memory is now given over to the monitor
 some processor time is consumed by the monitor

 Despite overhead, the simple batch system improves

utilization of the computer.
Multiprogrammed
Batch Systems
 Processor is
often idle
 even with
automatic
job
sequencing
 I/O devices
are slow
compared to
processor
 The processor spends a certain amount of
time executing, until it reaches an I/O
instruction; it must then wait until that I/O
instruction concludes before proceeding
 What if there’s enough memory to hold the OS (resident
monitor) and two user programs.
 When one job needs to wait for I/O, the processor can switch to
the other job, which may not be waiting.
 Multiprogramming
 also known as multitasking
 memory is expanded to hold three, four, or more programs
and switch among all of them
Multiprogramming
Example
Effects on Resource
Utilization

Table 2.2 Effects of Multiprogramming on Resource Utilization

Utilization Histograms
 Can be used to handle multiple interactive jobs
 Processor time is shared among multiple users
 Origin:multiple users simultaneously access
the system through terminals, with the OS
interleaving the execution of each user
program in a short burst or quantum of
computation
Table 2.3 Batch Multiprogramming versus Time Sharing
Compatible Time-Sharing
Systems
CTSS Time Slicing
 One of the first time-sharing  System clock generates interrupts at
operating systems a rate of approximately one every
0.2 seconds
 Developed at MIT by a group
known as Project MAC for IBM  At each interrupt OS regained
709/7094 control and could assign processor to
another user
 Ran on a computer with 32,000  At regular time intervals the current
36-bit words of main memory, with user would be preempted and
the resident monitor consuming another user loaded in
5000 of that
 Old user programs and data were
 To simplify both the monitor and written out to disk
memory management a program
was always loaded to start at the  Old user program code and data
location of the 5000th word were restored in main memory when
that program was next given a turn
 Operating
Systems are among the most
complex pieces of software ever developed
Major advances in
development include:
• Processes
• Memory management
• Information protection and
security
• Scheduling and resource
management
• System structure
 Fundamental to the structure of operating systems

A process can be defined as:

a program in execution
an instance of a running program

the entity that can be assigned to, and executed on, a processor

a unit of activity characterized by a single sequential thread of execution, a

current state, and an associated set of system resources
 Development of the
Process
 Three major lines of computer system development
created problems in timing and synchronization that
contributed to the development:
multiprogramming batch operation
• processor is switched among the various programs residing in main
memory

time sharing
• be responsive to the individual user but be able to support many users
simultaneously

real-time transaction systems

• a number of users are entering queries or updates against a database
 Causes of Errors
 Improper  Nondeterminate
synchronization program operation
 a program must wait until the  program execution is
data are available in a buffer interleaved by the processor
when memory is shared
 improper design of the
signaling mechanism can  the order in which programs
are scheduled may affect their
result in loss or duplication outcome

 Failed mutual exclusion  Deadlocks

 more than one user or  it is possible for two or more
program attempts to make programs to be hung up
use of a shared resource at waiting for each other
the same time
 may depend on the chance
 only one routine at at time
allowed to perform an timing of resource allocation
update against a given file and release
 A process contains  The execution context is
three components: essential:
 it is the internal data by
 an executable program which the OS is able to
 the associated data supervise and control the
process
needed by the program
 includes the contents of the
(variables, work space, various process registers
buffers, etc.)  includes information such as
the priority of the process and
 the execution context whether the process is waiting
(or “process state”) of for the completion of a
the program particular I/O event
Process
Management
 The entire state of the
process at any instant is
contained in its context

 New features can be

designed and incorporated
into the OS by expanding
the context to include any
new information needed to
support the feature
 The
OS has five principal storage
management responsibilities:

automatic
support of protection
process allocation long-term
modular and access
isolation and storage
programming control
management
A facility that allows programs to address
memory from a logical point of view, without
regard to the amount of main memory
physically available
 Conceived to meet the requirement of having
multiple user jobs reside in main memory
concurrently
 Allows processes to be comprised of a number of fixed-
size blocks, called pages
 Program references a word by means of a virtual address
 consists of a page number and an offset within the page
 each page may be located anywhere in main memory

 Provides for a dynamic mapping between the virtual

address used in the program and a real (or physical)
address in main memory
Virtual
Memory
Virtual Memory
Addressing
 The nature of the
threat that concerns
an organization will Main
vary greatly issues availability
depending on the
circumstances
authenticity confidentiality
 The problem involves
controlling access to data
computer systems integrity
and the information
stored in them
 Scheduling and
Resource Management
 Keyresponsibility of an OS is managing
resources
 Resource allocation policies must consider:

fairness
efficiency

differential
responsiveness
Key Elements of an
Operating System
 Different Architectural
Approaches
 Demands on operating systems require new
ways of organizing the OS

Different approaches and design elements have been tried:

• Microkernel architecture
• Multithreading
• Symmetric multiprocessing
• Distributed operating systems
• Object-oriented design
Microkernel Architecture
 Assigns only a few essential functions to the
kernel:
interprocess
address basic
communication
spaces scheduling
(IPC)

 The approach:

is well suited to a
simplifies provides
distributed
implementation flexibility
environment
 Technique in which a process, executing an application, is
divided into threads that can run concurrently

Thread
• dispatchable unit of work
• includes a processor context and its own data area to enable subroutine
branching
• executes sequentially and is interruptible
Process
• a collection of one or more threads and associated system resources
• programmer has greater control over the modularity of the application and the
timing of application related events
 Symmetric
Multiprocessing (SMP)
 Term that refers to a computer hardware architecture and also to the
OS behavior that exploits that architecture

 Several processes can run in parallel

 Multiple processors are transparent to the user

 these processors share same main memory and I/O facilities

 all processors can perform the same functions

 The OS takes care of scheduling of threads or processes on

individual processors and of synchronization among processors
 SMP Advantages
more than one process can be
Performance running simultaneously, each on a
different processor

failure of a single process does not

Availability halt the system

Incremental performance of a system can be

enhanced by adding an
Growth additional processor

vendors can offer a range of products

Scaling based on the number of processors
configured in the system
 Review/Reset
 Major OS Achievements:
 Processes
 Memory management
 Information protection, security
 Scheduling & resource management

 Developments Leading to Modern OSs

 Microkernel architectures
 Multithreading
 SMP/Distributed Operating Systems/OO Design
Distributed Operating Object-Oriented
System Design
 Provides the illusion of  Used for adding modular
 a single main memory space
extensions to a small kernel
 single secondary memory  Enables programmers to
space customize an operating system
 unified access facilities without disrupting system
integrity
 State of the art for distributed
operating systems lags that of
 Eases the development of
uniprocessor and SMP operating
systems distributed tools and full-blown
distributed operating systems
 Virtualization
 enables a single PC or server to simultaneously run multiple
operating systems or multiple sessions of a single OS
 Each (guest) operating system runs in a virtual machine (VM),
and can execute multiple applications
 Guest operating systems execute as if they were interacting
directly with the hardware, but in fact they are interacting
with a Virtual Machine Monitor (VMM) which runs
directly on the hardware or on a host operating
system
Virtual
Memory
Concept
Note: In some
cases, servers in
particular, the
VMM runs
directly on the
hardware.
This figure
represents a
hosted virtual
machine
Correction: should be Virtual MACHINE Concept
Process perspective:
• The machine on which it executes consists of
• the virtual memory space assigned to the process
• the processor registers it may use
• the user-level machine instructions it may execute
• OS system calls it may invoke for I/O
• ABI defines the machine as seen by a process

Application perspective:
• machine characteristics are specified by high-level language capabilities and OS
system library calls
• API defines the machine for an application

OS perspective:
• OS sees the system at the hardware level
• It (the OS) allocates real memory and I/O resources to the processes
• ISA provides the interface between the system and machine
 Process v System Virtual
Machine
 Process Virtual Machine  System Virtual Machine
 Is a virtual platform for  Each virtual machine is an
executing a single process isolated environment that
 Does not look like any real shares system resources with
hardware platform other VMs
 The process VM interprets  appears to the guest OS and
code that has been compiled the apps it runs as an exact
to run on the virtual machine copy of the real hardware
 Examples: JVM, Microsoft’s  The guest OS manages the
CLI (.NET) resources allocated to it by
the VMM
Process and System Virtual Machines
Process and System Virtual Machines
 Advantages
 Process Virtual Machine
 One compiler for multiple machines – generates byte code for JVM
 “Write once, run anywhere”
 .NET: Different modules of a program can be written in different
languages, all compile to the same format, can be compiled & run as a
unit

 System virtual machine

 Greater protection: esp. good for servers
 Multiple OS’s on one computer simultaneously
 Run legacy code on old versions of OS, run new programs on new OS
 An alternative to multiprogramming for resource sharing
 Symmetric Multiprocessor
OS Considerations
 A multiprocessor OS must provide all the functionality of a multiprogramming
system plus additional features to accommodate multiple processors

 Key design issues:

Simultaneous Scheduling Memory Reliability

concurrent Synchronization
management and fault
processes or tolerance
any
threads with multiple the reuse of
processor
may perform active processes physical
kernel routines having potential pages is the the OS
need to be scheduling,
which access to shared biggest should
reentrant to address spaces problem of provide
allow several complicates
the task of or shared I/O concern graceful
processors to resources, care degradation
execute the enforcing a
scheduling must be taken to in the face of
same kernel provide effective processor
code policy
synchronization failure
simultaneously
 Multicore OS
Considerations
 The design challenge for a
many-core multicore system is
to efficiently harness the hardware parallelism within
multicore processing power each core processor, known as
and intelligently manage the instruction level parallelism

substantial on-chip resources

efficiently potential for multiprogramming
and multithreaded execution
 Potential for parallelism exists within each processor
at three levels:
potential for a single application
to execute in concurrent
processes or threads across
multiple cores
 Developer must decide what pieces can or
should be executed simultaneously or in parallel
Grand Central Dispatch (GCD)

• implemented in Mac Os X 10.6

• helps a developer once something has been identified that
can be split off into a separate task
• thread pool mechanism
• allows anonymous functions as a way of specifying tasks
 Allows one or more cores to be dedicated to a
particular process and then leave the processor
alone to devote its efforts to that process
 Multicore OS could then act as a hypervisor that
makes a high-level decision to allocate cores and
some memory to applications but does little in
the way of resource allocation
beyond that.
 MS-DOS 1.0 released in 1981  Windows 2000
 4000 lines of assembly language  included services and functions to
source code
support distributed processing
 ran in 8 Kbytes of memory
 Active Directory
 used Intel 8086 microprocessor
 plug-and-play and power-
 Windows 3.0 shipped in 1990 management facilities
 16-bit
 Windows XP released in 2001
 GUI interface
 goal was to replace the versions of
 implemented as a layer on top of
Windows based on MS-DOS with
MS-DOS an OS based on NT
 Windows 95  Windows Vista shipped in 2007
 32-bit version
 led to the development of Windows
 Windows Server released in 2008
98 and Windows Me
 Windows 7 shipped in 2009, as well
 Windows NT (3.1) released in 1993 as Windows Server 2008 R2
 32-bit OS with the ability to support  Windows Azure
older DOS and Windows
applications as well as provide  targets cloud computing
OS/2 support
 Windows
Architecture
Kernel-mode components
1) Executive: contains core
system services
2) kernel: controls CPUs:
scheduling, process
switching, interrupts, etc.
3) HAL
4) Device drivers
5) Windowing, graphics
 Kernel-Mode Components
of Windows
 Executive
 contains the core OS services

 Kernel
 controls execution of the processors

 Hardware Abstraction Layer (HAL)

 maps between generic hardware commands and responses and those
unique to a specific platform

 Device Drivers
 dynamic libraries that extend the functionality of the Executive

 Windowing and Graphics System

 implements the GUI functions
 User-Mode Processes
 Four basic types are supported by Windows:
Special System • user-mode services needed to manage the system
Processes
• the printer spooler, event logger, and user-mode components that
Service Processes cooperate with device drivers, and various network services

Environment • provide different OS personalities (environments)

Subsystems
• executables (EXEs) and DLLs that provide the functionality users
User Applications run to make use of the system
 Windows OS services,  Advantages:
environmental subsystems,
and applications are all  it simplifies the Executive
structured using the  it improves reliability
client/server model
 it provides a uniform
 Common in distributed means for applications to
systems, but can be used communicate with
internal to a single system services via RPCs
without restricting
 Processes communicate flexibility
via RPC  it provides a suitable base
for distributed computing
 Two important characteristics of Windows are its support for
threads and for symmetric multiprocessing (SMP)
 OS routines can run on any available processor, and different routines can
execute simultaneously on different processors

 Windows supports the use of multiple threads of execution within a single

process. Multiple threads within the same process may execute on different
processors simultaneously

 Server processes may use multiple threads to process requests from more
than one client simultaneously

 Windows provides mechanisms for sharing data and resources between

processes and flexible interprocess communication capabilities
Windows Objects
 Windows draws heavily on the concepts of object-
oriented design
 Key object-oriented concepts used by Windows
are:
Object
Encapsulation class and Inheritance Polymorphism
instance
W
n
is
w
o
dK
rn
e
lC
o
O
tb
s
c
je
 Changes and improvements:
 Engineering improvements
 the system is now built in layers which can be separately tested
 Performance improvements
 amount of memory required has been reduced
 Reliability improvements
 user-mode heap is more tolerant of memory allocation errors by
C/C++ programmers
 Energy efficiency
 many improvements have been made
 Security
 BitLocker is now easier to set up and use
 Thread improvements
 can support hundreds of CPUs
 Dynamic Fair Share Scheduling (DFSS)
 Traditional UNIX Systems
 Were developed at Bell Labs and became operational on a PDP-7 in 1970

 Incorporated many ideas from Multics

 PDP-11was a milestone because it first showed that UNIX could be an OS for all
computers

 Next milestone was rewriting UNIX in the programming language C

 demonstrated the advantages of using a high-level language for system code

 Was described in a technical journal for the first time in 1974

 First widely available version outside Bell Labs was Version 6 in 1976

 Version 7, released in 1978 is the ancestor of most modern UNIX systems

 Most important of the non-AT&T systems was UNIX BSD (Berkeley Software
Distribution)
Description
of
UNIX
Traditional
UNIX
Kernel
Modern
UNIX
Kernel
 UNIX Versions
 System V, Release 4: (SVR4)
 A descendant of the original Bell Labs version
 Produced jointly by AT&T + Sun Microsystems
 One of the two main branches of UNIX
 BSD4.x (Berkeley Software Distribution 4.x)
 Many of the key developments in UNIX were
made at Berkeley sponsored by DOD
 BSD4.4 and FreeBSD are commonly used today
 Started out as a UNIX variant for the IBM PC

 Linus Torvalds, a Finnish student of computer science, wrote the initial

version

 Linux was first posted on the Internet in 1991

 Today it is a full-featured UNIX system that runs on several platforms

 Is free and the source code is available

 Key to success has been the availability of free software packages

 Highly modular and easily configured

 Modular
Monolithic Kernel
 Most UNIX systems are
Loadable Modules
monolithic (includes virtually all
of the OS functionality in one  Relatively independent blocks
large block of code that runs as a
single process with a single  A module is an object file whose
address space) code can be linked to and unlinked
from the kernel at runtime
 All the functional components
of the kernel have access to all  A module is executed in kernel
of its internal data structures mode on behalf of the current
and routines process

 Linux improves on this  Have two important

somewhat because it is characteristics:
structured as a collection of  Dynamic linking
modules
 Stackable modules
 Linux
Kernel
Modules
Linux Kernel Components
Linux Signals

Table 2.5 Some Linux Signals

 Linux Vserver Virtual
Machine Architecture
 Open-source, fast, lightweight
approach to implementing
chroot – UNIX or
virtual machines on a Linux Linux command to
chcontext –
make the root
server directory become
allocates a new
security context
something other
 Only a single copy of the than its default

Linux kernel is involved

 Supports a number of chbind – executes a
separate virtual servers command and
locks the resulting
capabilities – a
partitioning of the
 Each virtual server is isolated process and its
children into using
privileges available
to a root user
from the others a specific IP
address
 Involves four elements:
Linux Vserver Architecture
 Operating system objectives and  Process
functions:
 convenience, efficiency, ability to  Memory management
evolve  real address, virtual address
 user/computer interface
 Scheduling and resource management
 resource manager

 Multithreading
 Evolution:
 Symmetric multiprocessing (SMP)
 serial processing, simple batch  distributed OS
systems, multiprogrammed batch
 object oriented design
systems, time sharing systems
 Virtual machines
 Microsoft Windows/Windows 7
 virtualization
 UNIX/Linux systems