Operating Systems

Introduction

Chester Rebeiro
IIT Madras

Webpage : http://www.cse.iitm.ac.in/~chester/courses/15o_os/index.html
The Layers in Systems

Applications

Operating Systems

Computer
Organization

VLSI

Transistors

2
 OS usage

• Hardware Abstraction
turns hardware into something that
applications can use
• Resource Management
manage system’s resources

3
 A Simple Program
What is the output of the following program?

How is the string displayed on the screen?

4
 Displaying on the Screen
“Hello World” “Hello World” +
coordinates, color,
depth, etc

Monitor

Processor

Processor Memory
Graphics Card

• Can be complex and tedious

• Hardware dependent

Without an OS, all programs need to take care of every nitty gritty detail

5
 Operating Systems Provide
Abstraction
App

system call
(write to STDOUT)

Operating System Device

driver

• Easy to program apps

– No more nitty gritty details for programmers
• Reusable functionality
– Apps can reuse the OS functionality
• Portable
– OS interfaces are consistent. The app does not change when hardware changes

6
 OS as a Resource Manager
• Multiple apps but limited hardware
Apps

Operating Systems

A few processors
7
 OS as Resource Manager
• OS must manage CPU, memory, network,
disk etc…
• Resource management
– allows multiple apps to share resources
– protects apps from each other
– Improves performance by efficient utilization
of resources

8
 Sharing the CPU
App1 App2 App3 App4

Who uses the CPU?

App1 App2 App3 App4

time

pre9
 Operating Systems Types
• Application Specific
– Embedded OS
• eg. Contiki OS, for extremely memory constraint environments
– Mobile OS
• Android, iOS, Ubuntu Touch, Windows Touch
– RTOS
• QNX, VxWorks, RTLinux
– Secure Environments
• SeLinux, SeL4
– For Servers
• Redhat, Ubuntu, Windows Server
– Desktops
• Mac OS, Windows, Ubuntu

10
 JOS and xv6
• Designed for pedagogical reasons
• Unix like (version 6)
– Looks very similar to modern Linux operating
systems
• Theory classes : xv6
– Well documented, easy to understand
• Lab : JOS
– Build your own operating system from the
skeleton

11
 Course Structure
• Syllabus
– Overview of Operating Systems
– PC Hardware
– Memory Management
– Interrupts
– Context Switching
– Processes
– Scheduling
– Cooperating Processes
– Synchronization
– File Systems
– Security

12
 Textbooks / References
• ''xv6: a simple, Unix-like teaching operating system",
Revision 8, by Russ Cox, Frans Kaashoek, Robert
Morris
• ''Operating System Concepts'', 8th edition, by
Adraham Silberschatz, Pert B. Galvin, and Greg Gagne,
Wiley-India edition
• The xv6 source code booklet (revision 8)

http://www.cse.iitm.ac.in/~chester/courses/15o_os/index.html

13
 Logistics
• Theory Classes (CS24, Slot F)
– Wednesdays : 11:00 - 11:50 AM
– Thursdays : 9:00 - 9:50 AM
– Fridays : 8:00 - 8:50 AM
• Lab (System’s Lab, Slot P)
– Monday’s : 2:00 – 5:00 PMs

14
 Exams
• Quiz 1 : 25%
• Quiz 2 : 25%
• Final : 50%

15
Operating Systems

(How did it all start?)

16
 OS Evolution

• Evolution driven by Hardware improvements

+ User needs
– eg. low power requirements, Increased /
reduced security, lower latency
– Evolution by
• New/better abstractions
• New/better resource management
• New/better low level implementations

17
 Gen 1: Vacuum Tubes
• Hardware
– Vacuum tubes and IO with punchcards
– Expensive and slow
• User Apps
– Generally straightforward numeric computations done in machine
language

IBM Punch card

ENIAC 18
 Gen 1 : OS
• OS: Unheard of
• Human feeds program and prints output

George Ryckman, on IBM’s first computer

The cost of wastage was $146,000 per month (in 1954 US Dollars)

19
 Gen 2 : Mainframes
• Hardware
– transistors
• User Programs
– Assembly or Fortran entered using punch
cards
• OS : Batch systems
– Possibly greatest invention in OS
• Computers may be able to schedule their own
workload by means of software

20
 Batch Systems

• Operator collects jobs (through punch cards) and feeds it into a magnetic tape drive
• Special Program reads a job from input tape drive and on completion writes result to
output tape drive
• The next program is then read and executed
• Printing was done offline

21
 Batch Systems (pros.)
• Pros
– Better utilization of machine

22
 Batch Systems (cons.)
• In Batch Systems execute time includes reading
from input and writing to output.
• I/O considerably slower than execution
– Magnetic tapes were best read sequentially
• Therefore programmer must wait for long time

Input Output
Magnetic CPU Magnetic
Tape Tape

23
 Gen 3 : Mini computers
• Hardware
– SSI/MSI/LSI ICs
– Random access memories
– Interrupts (used to simulate concurrent execution)
• User Programs
– High level languages (Fotran, COBOL, C, …)
• Operating Systems
– Multiprogramming
– Spooling

24
 Multiprogramming
• Multiple jobs in memory
– When one waits for I/O the next job executes
• OS controls
– scheduling of jobs
– Protection between jobs
OS
Job 1 Memory
Job 2 partitions
Job 3

Multiprogramming with
3 jobs in memory

25
 Spooling
• Uses buffers to continuously stream inputs and outputs to
the system

Disk

Input Output
Magnetic Magnetic
Tape CPU Tape

• Pros : better utilization / throughput

• Cons : still not interactive

26
 Timesharing

Who uses the CPU?

Terminal 1 Terminal 2 Terminal 3 Terminal 4

time
27
 Timesharing
John McCarthy, 1962

28
 Multics, 1964
• Multiplexed Information and Computing Service
• Ambitious project started in MIT
• Introduced several new OS features but was not
successful by itself
– Segmented and Virtual memory
– High level language support
– Multi language support
– Security
– File system hierarchies
– Relational databases
– Shared memory multiprocessor

29
 Gen 4 : Personal Computers
• Hardware
– VLSI ICs
• User Programs
– High level languages
• Operating Systems
– Multi tasking
– More complex memory management and scheduling
– Synchronization
– Examples : Windows, Linux, etc

30
 Unix
• Dennis Ritchie and Ken Thomson tried to find an
alternative for Multics
• Appeared at the right time
Slater, 1987

31
 Unix adopted
• Spread and soon became widely adopted
Aho, 1984

• Wide spread adoption arguably a

hindrance to research?

32
 Smartphones & Tablets
• Hardware
– VLSI ICs, low power requirements & high compute
power
• Operating Systems
– User friendly
– Power awareness
– Always connected
– Offload to cloud
– Better protection, Virtual machines
– Examples : Android, iOS

33
 OS Buzzwords
• Buzzwords that have been around
Fairness
Security/
Utilization
Reliability

Isolation

• Contemporary buzzwords
multi core
support
application
energy / size
specific OS

virtualization

34
 OS Research Trends
Small
(footprint,
Minimum energy
requirements)

Security, Reliability Features

(fewer errors, (better device support,
Formally verified, Multi core support)
fault tolerant)

35