Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 51

AB1401

Information Technology

IT Infrastructure: Hardware

5.1 © 2009 by Pearson


LEARNING OBJECTIVES
 Define IT infrastructure and describe its
components
 Identify and describe the eras and technology
drivers of IT infrastructure evolution
 Assess contemporary computer hardware
platform trends

2
5.2 © 2009 by Pearson
ADDITIONAL READINGS
 Chip wars go to the Core
 Multicore
 Cloud Computing

3
5.3 © 2009 by Pearson
IT Infrastructure
 Defining IT infrastructure:
 Set of physical devices and software required to operate
enterprise
 Set of firmwide services including:
 Computing platforms providing computing services
 Telecommunications services
 Data management services
 Application software services
 Physical facilities management services
 IT management, standards, education, research and
development services
 “Service platform” perspective more accurate view of value
of investments

4
5.4 © 2009 by Pearson
Eras in IT Infrastructure
Evolution
 5 eras
 General-purpose mainframe and minicomputer era: 1959 to
present
 1958 IBM first mainframes introduced, eventually used to
support thousands of online remote terminals
 1965 less expensive DEC minicomputers introduced, allowing
decentralized computing
 Personal computer era: 1981 to present
 1981 Introduction of IBM PC
 Proliferation in 80s, 90s resulted in growth of personal software
 Client/server era: 1983 to present
 Desktop clients networked to servers, with processing work
split between clients and servers
 Network may be two-tiered or multitiered (N-tiered)
 Various types of servers (network, application, Web)

5
5.5 © 2009 by Pearson
Eras in IT Infrastructure
Evolution

Illustrated here are the typical computing configurations characterizing


each of the five eras of IT infrastructure evolution.

Figure 5-2A
6
5.6 © 2009 by Pearson
Eras in IT Infrastructure
Evolution
 Enterprise Internet computing era: 1992 to
present
 Move toward integrating disparate networks, applications
using Internet standards and enterprise applications
 Cloud Computing: 2000 to present
 Refers to a model of computing where firms and
individuals obtain computing power and software
applications over the Internet
 Fastest growing form of computing

7
5.7 © 2009 by Pearson
Eras in IT Infrastructure
Evolution

Figure 5-2B
8
5.8 © 2009 by Pearson
Eras in IT Infrastructure
Evolution
A Multitiered Client/Server Network (N-Tier)

In a multitiered client/server network, client requests for service are handled by different levels of servers.

Figure 5-3
9
5.9 © 2009 by Pearson
Technology drivers of
infrastructure evolution
 Moore’s law and microprocessing power
 Computing power doubles every 18 months
 Nanotechnology: May shrink size of transistors to
width of several atoms
 Contrary factors: Heat dissipation needs, power
consumption concerns
 Law of Mass Digital Storage
 The amount of data being stored each year
doubles

10
5.10 © 2009 by Pearson
Technology drivers of
infrastructure evolution
Moore’s Law and Microprocessor Performance

Packing more transistors into a tiny microprocessor has exponentially increased processing power.
Source: 2004 Intel Corporation; updated by the authors.

Figure 5-4
11
5.11 © 2009 by Pearson
Technology drivers of
infrastructure evolution
Falling Cost of Chips

An Intel® processor today


can contain as many as 1
billion transistors, run at 3.2
GHz and higher, deliver over
10,000 MIPS, and can be
manufactured in high
volumes with transistors
that cost less than
1/10,000th of a cent. That’s a
little less than the cost of
one printed character in this
book.

Figure 5-5
12
5.12 © 2009 by Pearson
Technology drivers of
infrastructure evolution
The Capacity of Hard Drives Grows Exponentially
1980-2007

From 1980 to 1990, hard disk drive capacities for PCs grew at the rate of 25 percent annual compound growth, but after 1990,
growth accelerated to more than 65 percent each year.

Figure 5-7
13
5.13 © 2009 by Pearson
Technology drivers of
infrastructure evolution
The Cost of Storing Data Declines Exponentially
1950-2010

Since the first magnetic


storage device was used in
1955, the cost of storing a
kilobyte of data has fallen
exponentially, doubling the
amount of digital storage
for each dollar expended
every 15 months on
average.

Figure 5-8
14
5.14 © 2009 by Pearson
Technology drivers of
infrastructure evolution
 Metcalfe’s Law and network economics
 Value or power of a network grows
exponentially as a function of the number
of network members
 As network members increase, more
people want to use it (demand for network
access increases)

15
5.15 © 2009 by Pearson
Technology drivers of
infrastructure evolution
 Declining communication costs and the
Internet
 An estimated 1.5 billion people worldwide
have Internet access
 As communication costs fall toward a very
small number and approach 0, utilization of
communication and computing facilities
explodes

16
5.16 © 2009 by Pearson
Technology drivers of
infrastructure evolution
Exponential Declines in
Internet Communication Costs

One reason for the growth in the Internet population is the rapid decline in Internet connection and overall communication costs.
The cost per kilobit of Internet access has fallen exponentially since 1995. Digital Subscriber Line (DSL) and cable modems now
deliver a kilobit of communication for a retail price of less than 2 cents.
Figure 5-9
17
5.17 © 2009 by Pearson
Technology drivers of
infrastructure evolution
 Standards and network effects
 Technology standards:
 Specifications that establish the compatibility of
products and the ability to communicate in a
network
 Unleash powerful economies of scale and result
in price declines as manufacturers focus on the
products built to a single standard

18
5.18 © 2009 by Pearson
Different types of computers
 PDA, handheld computers, Tablet PC (Palm, HP-
Compaq, O2, Dopod)
 Nettop,Netbook (fit-PC, CherryPal, Koolu, Mac Mini,
ASUS Linutop, Zonbu, gPC and gPC mini by Everex)
 Laptop/Notebook computer (Toshiba, Fujitsu, Lenovo
(formally IBM))
 Desktop computer (Lenovo, HP-Compaq)
 Workstations (Sun, Silicon Graphics)
 Minicomputers/mainframes (IBM, HP-Compaq)
 Supercomputers (Cray, NEC, Fujitsu)

19
5.19 © 2009 by Pearson
Computer hardware platforms
 Computer hardware platforms
 Client machines
 Desktop PCs, mobile computing devices – PDAs, laptops,
Nettops, Netbooks, etc
 Servers
 Blade servers: ultrathin computers stored in racks
 Mainframes:
 IBM mainframe equivalent to thousands of blade servers
 Top chip producers: AMD, Intel, IBM
 Top firms: IBM, HP, Dell, Sun Microsystems

20
5.20 © 2009 by Pearson
Computer hardware platforms
 Components of a computer
 Input devices
 Processing unit
 Output devices
 Storage devices
 Communications devices

21
5.21 © 2009 by Pearson
Computer hardware platforms
 Input devices
 Different devices may be preferred for
different applications
 More? (video: multi-touch computing)

22
5.22 © 2009 by Pearson
Computer hardware platforms
 Processing unit (CPU and memory)
 Most PCs use CPUs made by Intel or AMD

 Common memory format today is DDR SDRAM


 double-data-rate synchronous-dynamic-random-
access-memory (what a mouthful!)

23
5.23 © 2009 by Pearson
Computer hardware platforms
 Processing unit (CPU and memory) (cont.)
 The CPU and memory sit in their sockets on the
motherboard

24
5.24 © 2009 by Pearson
Computer hardware platforms
 Output devices
 Output devices produce information in
human-accessible modes

25
5.25 © 2009 by Pearson
Computer hardware platforms
 Storage devices
 Storage media may be fixed or removable,
magnetic or optical
 More?

26
5.26 © 2009 by Pearson
Computer hardware platforms
 Communication devices
 Communication devices may be wired or
wireless

27
5.27 © 2009 by Pearson
Computer hardware platforms

28
5.28 © 2009 by Pearson
How a CPU Works?

5.29p. 6.170 Fig. 6.9 © 2009 by Pearson


How a CPU works, cont.

1. Retrieve
4.
4. Store
Store the
the result an
In
In RAM
RAM instruction
from RAM

Machine cycle consists of:

3.
3. Execute
Execute the
the 2.
2. Decode
Decode the
instruction
instruction instruction

5.30 © 2009 by Pearson


Computer hardware platforms
 Data are stored in …
1. Level 1 cache memory (on chip)
2. Level 2, Level 3 cache memory (on chip
or motherboard)
3. Random Access Memory (RAM)
4. Hard disk
5. Other storage devices (CD, floppies,
thumbdrive, …)
6. Other computers (over the network)

31
5.31 © 2009 by Pearson
Searching the memory hierarchy
Progressive search
down levels
Found in
NO L1 cache

Found in
L2 cache
NO
Found in
RAM
NO
Found
5.32 onPearson
© 2009 by disk
Computer hardware platforms
 Fetching data to the CPU
 Level 1 cache memory (1-10 nanoseconds)
 Level 2, 3 cache memory (10-50 nanoseconds)
 Random Access Memory (50-500 nanoseconds)
 Hard disk (10 milliseconds)
 CD-ROM / floppy (100 millisecond-1 second)
 Network response times (many seconds)
* nanosecond=10-9 second, millisecond=10-3 second

33
5.33 © 2009 by Pearson
Computer hardware platforms
 Implications of the model
 L1, L2, L3 caches helps a lot
 Get more RAM
 Disk access is a disaster for performance
 If data resides on media slower than hard
disk (eg. CD / floppy), copy it to hard disk
 Copy network files to hard disk before
opening

34
5.34 © 2009 by Pearson
Computer hardware platforms
 Computer performance
 Do you judge a computer by the clock speed (in
GHz) of its CPU?
 This is like judging cars on the basis of their
engine horsepower
 ignores other factors that have an effect on performance
(gears, wheels, weight, tyres, …)
 CPU speed is just one of many factors affecting
computer performance
 RAM (size, speed)
 Bus width (width, speed)

35
5.35 © 2009 by Pearson
Clock Speed

CPU
The CPU executes a series of instructions one at a time
as dictated by a program. The speed of a CPU is
measured by how many instructions it can perform in
one second. Clock rate, measured in gigahertz, and
available cache memory determine how quickly the
CPU can perform instruction cycles of fetch, decode
and execute.

5.36 © 2009 by Pearson


Clock Speed

Each beat or tick of the


clock is called a CPU
cycle/machine cycle CPU speed is quoted in
Megahertz (MHz = 1 million
CPU cycles per second) or
Gigahertz (GHz= 1 billiion
CPU cycles per second).

5.37 © 2009 by Pearson


Bus Speed
Bus
The buses of a computer are the routes in which data is
transmitted in the computer. Local buses connect
hardware components in a computer system, while
external buses connect the computer to external
peripherals. The speed of a bus is related to its
proximity to the CPU and connection to cache memory.

Front-Side Bus
The front-side bus is the computer's central route for
receiving and transmitting data to the hardware of the
computer through the memory controller hub and the
input/output controller hub. The front-side bus is not
directly connected to the cache and thus operates at
only a fraction of the CPU's clock rate.
5.38 © 2009 by Pearson
Bus Speed
Backside Bus
The backside bus refers to the connection between the
CPU and the cache memory. Because the backside bus
resides on the same die as the chip, it can operate at or
near the full clock rate of the CPU.

PCI and AGP Buses


The peripheral component interconnect (PCI) and
accelerated graphics port (AGP) are buses that connect
the computer to peripherals. PCI shares bandwidth--the
limit of how much information can travel along a bus--
with multiple devices, while AGP is dedicated to graphic
cards.

5.39 © 2009 by Pearson


Computer hardware platforms
 Multicore processors
 Dual-core, quad-core laptops already in the market
 Future: 8-core, 16-core, …
 Benefits of multi-core processors:
 Cost savings by reducing power requirements and hardware sprawl
 Less costly to maintain as fewer systems need to be monitored
 Performance and productivity benefits beyond the capabilities of today’s
single-core processors
 Able to handle the exponential growth of digital data and the
globalization of the Internet
 Able to meet the demands of sophisticated software applications under
development
 Run applications more efficiently than single-core processors – giving
users the ability to keep working even while running the most
processor intensive task in the background
 Able to increase performance in areas such as data mining,
mathematical analysis, and Web serving

40
5.40 © 2009 by Pearson
Processor, Core, Thread
 A CPU typically has only one Core
 Instead of having multiple CPUs, multiple
Cores are built into one CPU instead
 Such are known as multiprocessor computers
 Only high-end data servers, and supercomputers
would have need for multiple CPUs
 Hyper threading
 A feature of certain Intel chips that makes one
physical CPU / Core appear as two logical CPUs /
Cores
41
5.41 © 2009 by Pearson
42
5.42 © 2009 by Pearson
Running Computer Tasks
 Multitasking, in an operating system, is
allowing a user to perform more than one
computer task at a time.
 When you open your Web browser and

then open Word at the same time, you are


causing the operating system to do
multitasking.
 Multitask on single core CPU

 Multitask on multicore CPU aka

multiprocessing

43
5.43 © 2009 by Pearson
Multitasking in Single Core CPU

5.44 © 2009 by Pearson


Multitasking in Multicore CPU

5.45 © 2009 by Pearson


Running Computer Programs
 Parallel processing is
 the processing of program instructions by dividing
them among multiple processors / cores with the
objective of running a program in less time.
 The simultaneous use of more than one CPU to
execute a program
 Multiprogramming is a form of parallel
processing in which several programs are run
at the same time on a processor / core
 Since there is only one processor, there can be no true
simultaneous execution of different programs. Instead, the
operating system executes part of one program, then part
of another, and so on. To the user it appears that all
programs are executing at the same time.
46
5.46 © 2009 by Pearson
47
5.47 © 2009 by Pearson
Group Discussion
 “Tablets, from the iPad to Kindle to RIM’s
Playbook, are all a by-product of cloud
computing.” Explain if you agree or
disagree with this statement.

(November/December 2011 Exam Question)

5.48 © 2009 by Pearson


Group Discussion
 Why are standards so important in
technology? Give two examples of
technology standards and explain the
significance of any one standard. Also
highlight the consequences if it did not
exist.
(April/May 2013 Exam Question)

5.49 © 2009 by Pearson


Class Discussion
 Does the size of RAM affect the
performance of your computer? How?

5.50 © 2009 by Pearson


Computer hardware platforms
 Discussion
 Does a dual-core computer perform at
twice the speed of a single-core computer?
(Hint: Reading - Chip wars go to the core)
 The more core the merrier? (Hint: Reading
– Multicore)

51
5.51 © 2009 by Pearson

You might also like