Information and Communication Technology Policy and Strategy
sml0om7
01wkict17.pptxThe Waves of Change
1
Introducing your Lecturer – Ed Steinmueller
Ed has a PhD in economics from Stanford where he wrote a thesis on the integrated circuit industry. He studies at Stanford and worked in the Palo Alto, California area for 20 years before emigrating to Europe.
His work involved an economic policy research at Stanford (CEPR now called SIEPR) and a partnership providing expert testimony in competition policy legal disputes (involving newspapers and a variety of high technology industries including telecommunications, minicomputers, and nuclear fuels)
He came to Europe in 1994 and worked in Maastricht at MERIT, a sister institution to SPRU where he founded a PhD programme and began work on European information society issues. In 1997 he came to SPRU and has continued his work on social and economic issues surround ICTs.
Most recently, he has been working on issues of social, technical and economic transition to more sustainable societies.
Outline of Lecture
Technological Revolution
Technological Determinism
Innovation Diffusion
The Productivity Paradox
3
Technological Revolution
The modern ICT (information and communication technology) era is marked by four basic trends
The integrated circuit paradigm of Moore’s Law
The application of optoelectronics to communication
High capacity mass storage
Increasing mobility of access to ICT
And two important ‘standards’ innovations
Pack switching giving rise to the Internet Protocol
World Wide Web standards (HTML, URL, etc.)
These trends reflects the power of
a) ‘technological trajectories’ -- the rapid incremental improvement of performance and capacity.
b) ‘standards’ that provide the basis for interconnection and network externalities (which we will discuss in more detail in Week 4)
4
1. The Integrated Circuit Paradigm
Pace of Change (Moore’s Law)
The increase in the number of transistors per device has doubled every 18 months
since 1960. The persistence of this pace is remarkable and has few parallels in
other technologies.
Consequences of Moore’s Law
1. Miniaturisation (not primarily about making things smaller)
Reductions in interconnections
Reductions in power consumption
2. Increases in functionality because
a) integration involves ‘digestion’ of all previous electronic designs into single ICs
b) as well as the capacity to evolve ever more complex and general purpose ‘systems on a chip’
Technological Revolution
5
Moore’s Law Illustrated:
The number of transistors has (more than) doubled every 18 months since 1960
(depicted on a logarithmic scale)
Source: Intel
1. The Integrated Circuit Paradigm
Technological Revolution
6
1 x 1.6 mm Texas Instruments First Integrated Circuit (Jack Kilby)
One transistor, one capacitor and the equivalent of 3 resistors
1958
a. The Integrated Circuit Paradigm
Technological Revolution
Inventors of the IC
Jack Kilby Robert Noyce
7
1965
15 x 15 mm Fairchild Operational Amplifier
14 transistors and fifteen resistors
a. The Integrated Circuit Paradigm
Technological Revolution
8
0.09 x 0.11 mm Section of the AMD 2901 1975 – An Early Microprocessor
c. 5,000 transistors in the entire device
1975
a. The Integrated Circuit Paradigm
Technological Revolution
9
a. The Integrated Circuit Paradigm
2004
This is the Pentium 4 ‘Prescott’ design which has 125 million
transistors.
It is no longer possible to display the detail of a modern IC because the number of pixels on a screen is far fewer than the number of transistors (a computer screen with 1920x1080 resolution only has about 2 million pixels <1% of the number of transistors)
A picture of a modern IC is like a satellite image of London
A nice short (15 minute) lecture on small scale is
https:// nanohub.org/resources/179/watch?resid=20231&tmpl=component&time=00:14:09
Technological Revolution
The Emergence of the Integrated Circuit Paradigm (1)
Why was the military the first user of ICs?
ICs were initially much more expensive than the costs of buying the transistor and assembling them on a printed circuit board.
The cost premium could only be justified in missile and other aerospace applications where higher standards of reliability demanded fewer interconnections.
How did commercial use emerge?
1) more complex devices (hearing aid amplifiers, calculators) became possible as
lithography techniques for scaling to smaller dimensions improved in quality and
became less expensive -- note that there is an element of physical miniaturisation
in these early applications
2) by the late 1960s the ability to ‘integrate’ a larger number of transistors meant that ICs became cheaper than transistors assembled on printed circuit boards
Price dynamics
Early years – yields were low and high prices reflected manufacturing costs
Later – high demand for early applications support high prices (and prevent demand for earlier generation from collapsing)
a. The Integrated Circuit Paradigm
Technological Revolution
11
The Emergence of the Integrated Circuit Paradigm (2)
As the number of circuit elements increased ever more complex systems
could be ‘integrated’ on a single circuit
This process has now accelerated so that modern integrated circuits are constructed from libraries of circuit elements – some of these library elements are much more complex than early integrated circuits
A key feature of the paradigm is the principle of ‘modularity’ – very similar to software modularity in which circuit elements have defined inputs and outputs that facilitate their integration into larger systems
A consequence is that, in principle, any software can be implemented as hardware
An integrated circuit design can be accurately simulated using software so that very complex systems can be tested before manufacture
Moore’s law is not only a prediction, it is a roadmap guiding technological development in the industry
a. The Integrated Circuit Paradigm
Technological Revolution
12
b. The Application of Opto-Electronics
Opto-electronics is the integration of digital electronics and lasers
Lasers can be switched at high rates of speed to produce streams of digital data
Like electricity, light travels at the speed of light – the time to circle the earth at the equator is a bit more than 1/10 of a second
(In practice, signals move slower because they must be ‘switched’ (recognised and then re-sent))
Packet switching over fibre-optic cables allows global data and voice communication at very low (and falling) costs
The essence of the Internet is packet switched data transmission which allows the volume of data traffic (now in millions of terabytes per month)
A technical introduction to Fibre Optics is available at http://fiberu.org/
Technological Revolution
13
c. High capacity mass storage
Magnetic and optical mass storage densities and speeds have increased dramatically since
This is remarkable because such devices have ‘moving parts’ – they only partly benefit from Moore’s law
High capacity mass storage underlies the server technologies responsible for the World Wide Web
In recent years, solid state memories have begun to rival mechanical mass storage for some applications – e.g. the USB ‘key’ and increasingly SSD (solid state disks)
This is an example of how a linear (but rapid) growth (in hard disk capacity) may be displaced by exponential growth (in memory chips)
Technological Revolution
14
"Full History Disk Areal Density Trend" by Barry Whyte. Licensed under CC BY 4.0 via Wikimedia Commons - http://commons.wikimedia.org/wiki/File:Full_History_Disk_Areal_Density_Trend.png#mediaviewer/File:Full_History_Disk_Areal_Density_Trend.png
Hard Disk Capacity Follows (and even exceeds) Moore’s Law
Cellular networks make it possible to use the limited capacity of the radio
Frequency (RF) spectrum to carry data as an RF signal.
Central to the RF technological trajectory is the concept of cellular networks
From a technical perspective the amount of data that can be exchanged in such a network is limited by radio frequency with higher radio frequencies suffering from problems of interference (e.g. rain) and attenuation (e.g. reduction in signal while penetrating solid objects)
These limitations can be addressed by reducing the power and hence shortening the effective range of the signal. It is also helpful to be able to shift the frequency to avoid congestion and improve signal quality.
The use of RF for ICT applications depends upon ‘frequency allocation’ – rules that determine what legal use can be made of the RF spectrum (see next slides) and the limitations in this frequency allocation require inventing techniques to more fully utilise the allocated frequency (e.g. frequency shift)
d. Mobility Using RF Networks
Technological Revolution
Wifi 2.4 & 5 GHz
WiMax 2-66 GHz
(mostly 2.3 & 2.5 GHz)
Mobile
Many frequencies including
925 MHz, 1.8 GHz,
And 1.93 GHz
Technological Revolution
d. Mobility Using RF Networks
Technological Revolution
The RF spectrum is crowded and mobile RF applications (purple) have had
to fit into prior allocation to other uses (e.g. broadcasting and
(not depicted here) military applications
d. Mobility Using RF Networks
Heddy Lamarr
(Hedwig Eva Maria Kiesler)
(1914-2000)
George Antheil
(1900-1959)
Frequency hopping (shifting) was invented by these two individuals as a contribution to the Allies efforts during WW II and is now widely employed to reduce ‘congestion effects’ in RF frequencies. While the most glamourous, these were only two of dozens of important inventors of RF technologies.
Technological Revolution
d. Mobility Using RF Networks
Unifying Principle – Technological Trajectories
A technological trajectory is a projection or forecast of performance and capacity
As is the case with Moore’s law the forecast provides a means of organising and directing technological research
None of the three fundamental technologies responsible for the ICT revolution have encountered fundamental roadblocks – this is unusual
They are all related to a single body of scientific knowledge – materials science
The same area of science which is now opening up nanotechnology for development
There are also important connections between materials science and biotechnology
Technological Revolution
20
Technological Determinism (1)
What I have just presented is a version of ‘technological determinism’
I have suggested that forecasts of improvement guide efforts to make improvement, creating a virtuous cycle whose consequences are the ICT revolution
This view is contested by social scientists who observe that technological efforts are planned, directed, and carried out by people who influence the rate and direction of technological improvement – they are ‘socially constructed’
There is certainly some truth in this argument and some relevance for our purposes
In a competitive market economy, technological opportunities not exploited by one group of people will be exploited by others
A consequence of the absence of roadblocks to technological advance in the three central technologies is that a global technological race has been underway for several decades – it shows no signs of decline
This race has more to do with ‘raw’ technological capacity than it does with the use of the technology because production improvements are an easier target than use improvements (for reasons we will examine further in this and the next classes)
21
Technological Determinism (2)
It is true, however, that the ways in which technology is employed are influenced by social groups such as the technical staff of ICT producers and the users of ICT
An important part of the discussion of this course is considering who gets included and excluded in these processes of ‘social construction’ of technology
For example, one of the key features of ICT developments is that races to exploit technological opportunity often produce dominant companies
WINTEL -- Microsoft and Intel
CISCO
SAP and Oracle
Nokia and Motorola and now Samsung and Apple
22
Technological Determinism (3)
Information and communication technologies (hereafter ICTs) are subject to several types of economies of scale:
Many ICTs have important elements of fixed costs – e.g. the software example.
The ‘economy of integration’ from the IC industry makes it possible to produce ever more complex devices at similar costs – thus, over time the ‘real’ (quality adjusted) price of electronics falls which increases demand
Investments in design capabilities can be applied to a number of different areas (economies of scope)
Many ICTs are mass produced (e.g. optical disk drives) and there are classical mass production economies of scale available.
In addition, they are affected by:
1) Adoption Externalities – what others do influences your choice
2) Technological Standards (week 4) in augmenting or strengthening economies of scale.
23
Innovation Diffusion (1)
24
The ‘Stylised Facts’ of the Diffusion Curve
I. Early Adoption
II. Take Off
III. Rapid Growth
IV. Maturity
V. Obsolescence
Innovation
I
II
III
IV
V
Share
of Total
Available
Market
Time
Innovation Diffusion (2)
25
Innovation Diffusion (3)
Consequences of Diffusion
The diffusion curve continues the ‘technological determinist’ perspective
It suggests that new technologies are necessarily taken up according to the ‘stylised facts’ based upon experience
The rapid rate of the ‘take up’ phase in combination with economies of scale are responsible for the ‘winner takes all’ nature of technological races
But are ‘diffusion models’ very good at prediction?
26
Innovation Diffusion (4)
Predicting Diffusion (a)
27
Innovation Diffusion (5)
Predicting Diffusion (b)
28
Another Prediction
From about 1960, social commentators have been talking about ‘automation’ as a potential threat to jobs and as a remarkable new source of productivity that might enable changes in the service industry comparable to those achieved through electrical power in manufacturing
With some conspicuous exceptions (e.g. telephone operators) automation has not produced the extent of job losses or the productivity improvements predicted.
Why not?
The principal reason is that ICTs support the creation of variety (product differentiation) as effectively as they improve the control of mass production and consumption -- The new opportunities for growth obscure the ‘disruption’ and ‘displacement’ effects enabled by ICTs
A secondary reason is that the expectations that the ‘conversion’ process represented by automation would be smooth or automatic are consistently frustrated – organisational change is difficult – i.e. it is costly and time-consuming
The Productivity Paradox
29
Management Challenges
Managing the ‘disruptive’ elements of change requires a better understanding of how and why disruption occurs in the process of adopting ICT solutions
There is no smooth path between how a system may be designed according to the ideas of the technologist designing it and and how it operates in practice as part of a social system
Central to this better understanding is an appreciation of how ICTs enter into the social systems in which they are expected to have a role
These issues are dealt with in this course by considering engagement, co-ordination and integration issues (statements of ideas that link technological and social science ways of thinking)
At both the level of the firm and in society as whole – ICTs are seen as a progressive force
There are limits and problems with this viewpoint and understanding what these are provides better insight into how ICT take up and use can be managed
The Productivity Paradox
30
Technological Determinism (Soft and Hard) vs. Social Constructivism
When we talk about the ICT revolution, who or what are the revolutionaries?
‘Hard’ technological determinism is the view that technology is autonomous of direct human control and dictates changes in society
Social constructivism rejects technological determinism because it
conceals power relationships
ignores the human instrumentality in creating technology and human choices made in implementing technology, and
provides little basis for understanding why technologies fail to deliver the promises of their supporters
Summary (1)
Social constructivism has its own risks, however. It is too easy to see everything as contingent on (depending on) power relationships and ignore the influence of technological opportunities on choices.
Soft technological determinism recognises that human agency plays a major role in technological outcomes, but maintains that technological opportunities are not solely socially constructed and that the variety of initiatives by actors produces ‘emergent outcomes’ whose path is not controlled by any actor or group of actors
31
Relevant Sociological and Political Ideas
The social constructivist debate described earlier indicates the importance of examining who makes the decisions regarding the design of ICT systems
-- Often this is characterised as the contest between ‘supply push’ and ‘demand pull’ types of explanations, but this obscures the point as much as it illuminates it
-- Defining what social interests there are in ICTs is an important subject
At the level of the individual user, issues of ‘resistance’ and ‘accommodation’ are important
-- ICTs make existing skills obsolete and demand new skills
ICTs are also a new method for achieving ‘co-ordination’ between individuals in an organisation – co-ordination occurs when we must take account of others actions.
-- For example, express parcel delivery
-- The larger implications of these new means of ‘co-ordination’ are only beginning to be recognised (e.g. open source software movement).
Summary (2)
32
Looking Forward
Some key questions for the module:
How are people either left behind or excluded by ICT developments – do digital divides exist? If yes, what are their consequences.
Why is technological change faster than organisational change?
3. Is improvement possible or is this another instance of ‘determinism’
Beginning with the technological opportunity reveals my own bias toward a soft technological determinism approach. For me, it all begins with the remarkable potential of the technology.
The ‘bridge’ or ‘road’ to the Information Society is not, however, likely to be smooth due to both economic and sociological/political considerations.
The next lecture continues the use of the economic dimension by looking at the role of the Internet in markets which is followed by the guest lecture of Dr. Puay
Tang who will look at a key rule (institution) – intellectual property rights. In week 4 we will examine another institution (rule) – standards and how they influence ‘connectivity’ (the formation of networks) and then discuss the implications of networks in the following lectures
33
(c) Frank M. Bass (1999)
Projection of World
-
Wide PC Demand, 1999
-
2010
-
Data From Bill Gates,
Newsweek
Actual Worldwide PC Shipments, 1981-1999 and Fitted and Projected
Shipments, 1981-2010, m=3.384 Billion, p= .001, q= .195
0
20
40
60
80
100
120
140
160
180
81
83
85
87
89
91
93
95
97
99
101
103
105
107
109
Year
Millions of Units
World Wide PC Shipments
Fitted World Wide PC Shipments
Peak
2008
697 Million Units
Shipments through
1999
Shipments Includes Replacements
(Upgrades)
5
-
31
-
99
Projection of World-Wide PC Demand, 1999-2010-Data From Bill Gates, Newsweek
5-31-99
(c) Frank M. Bass (1999)
Chart1
| 81 |
| 82 |
| 83 |
| 84 |
| 85 |
| 86 |
| 87 |
| 88 |
| 89 |
| 90 |
| 91 |
| 92 |
| 93 |
| 94 |
| 95 |
| 96 |
| 97 |
| 98 |
| 99 |
Chart2
| 81 | 81 |
| 82 | 82 |
| 83 | 83 |
| 84 | 84 |
| 85 | 85 |
| 86 | 86 |
| 87 | 87 |
| 88 | 88 |
| 89 | 89 |
| 90 | 90 |
| 91 | 91 |
| 92 | 92 |
| 93 | 93 |
| 94 | 94 |
| 95 | 95 |
| 96 | 96 |
| 97 | 97 |
| 98 | 98 |
| 99 | 99 |
Chart3
| 81 | 81 |
| 82 | 82 |
| 83 | 83 |
| 84 | 84 |
| 85 | 85 |
| 86 | 86 |
| 87 | 87 |
| 88 | 88 |
| 89 | 89 |
| 90 | 90 |
| 91 | 91 |
| 92 | 92 |
| 93 | 93 |
| 94 | 94 |
| 95 | 95 |
| 96 | 96 |
| 97 | 97 |
| 98 | 98 |
| 99 | 99 |
| 100 | 100 |
| 101 | 101 |
| 102 | 102 |
| 103 | 103 |
| 104 | 104 |
| 105 | 105 |
| 106 | 106 |
| 107 | 107 |
| 108 | 108 |
| 109 | 109 |
| 110 | 110 |
Sheet1
| Year | World Wide PC Shipments | Time | World Wide PC Shipments | Fitted World Wide PC Shipments | ||||
| 81 | 2.75 | 0 | 2.75 | 3.384 | ||||
| 82 | 5.49 | 1 | 5.49 | 4.1076463616 | ||||
| 83 | 6.53 | 2 | 6.53 | 4.9836726758 | ||||
| 84 | 12.5 | 3 | 12.5 | 6.0430441788 | ||||
| 85 | 10.91 | 4 | 10.91 | 7.3224876365 | ||||
| 86 | 13.73 | 5 | 13.73 | 8.865309973 | ||||
| 87 | 16.47 | 6 | 16.47 | 10.722206607 | ||||
| 88 | 19.22 | 7 | 19.22 | 12.9519810364 | ||||
| 89 | 21.96 | 8 | 21.96 | 15.6220536671 | ||||
| 90 | 24.705 | 9 | 24.705 | 18.8085791372 | ||||
| 91 | 27.45 | 10 | 27.45 | 22.5959157478 | ||||
| 92 | 30.195 | 11 | 30.195 | 27.0750992052 | ||||
| 93 | 46.47 | 12 | 46.47 | 32.3408724454 | ||||
| 94 | 51.96 | 13 | 51.96 | 38.4867305399 | ||||
| 95 | 57.45 | 14 | 57.45 | 45.5973867868 | ||||
| 96 | 65.49 | 15 | 65.49 | 53.7381064892 | ||||
| 97 | 87.45 | 16 | 87.45 | 62.9405660702 | ||||
| 98 | 95.49 | 17 | 95.49 | 73.1853723349 | ||||
| 99 | 100.98 | 18 | 100.98 | 84.3822087031 | ||||
| 100 | 19 | 96.3497923986 | ||||||
| 101 | 20 | 108.7993185126 | Raw R-square (1-Residual/Total) = 0.996 | |||||
| 102 | 21 | 121.3264896586 | Mean corrected R-square (1-Residual/Corrected) = 0.990 | |||||
| 103 | 22 | 133.4179599354 | R(observed vs predicted) square = 0.990 | |||||
| 104 | 23 | 144.4772493566 | Bass Model Estimates | |||||
| 105 | 24 | 153.8722322115 | Wald Confidence Interval | |||||
| 106 | 25 | 161.0011415572 | Parameter Estimate A.S.E. Param/ASE Lower < 95%> Upper | |||||
| 107 | 26 | 165.3677583496 | M 3384.069 1303.840 2.595 620.052 6148.086 | |||||
| 108 | 27 | 166.6512970306 | P 0.001 0.000 3.795 0.001 0.002 | |||||
| 109 | 28 | 164.7550347272 | Q 0.195 0.017 11.353 0.158 0.231 | |||||
| 110 | 29 | 159.8214777875 | ||||||
| 697.2 |
Sheet2
Sheet3
Actual Worldwide PC Shipments, 1981-1999 and Fitted and Projected
Shipments, 1981-2010, m=3.384 Billion, p= .001, q= .195
0
20
40
60
80
100
120
140
160
180
81838587899193959799
101103105107109
Year
Millions of Units
World Wide PC ShipmentsFitted World Wide PC Shipments
Peak
2008
697 Million Units
Shipments through
1999
Shipments Includes Replacements
(Upgrades)
Actual vs. Bass Predicted PC Shipments
0
50
100
150
200
250
300
350
19751980198519901995200020052010
Year
Millions Shipped
Chart1
| 1981 | 1981 |
| 1982 | 1982 |
| 1983 | 1983 |
| 1984 | 1984 |
| 1985 | 1985 |
| 1986 | 1986 |
| 1987 | 1987 |
| 1988 | 1988 |
| 1989 | 1989 |
| 1990 | 1990 |
| 1991 | 1991 |
| 1992 | 1992 |
| 1993 | 1993 |
| 1994 | 1994 |
| 1995 | 1995 |
| 1996 | 1996 |
| 1997 | 1997 |
| 1998 | 1998 |
| 1999 | 1999 |
| 2000 | 2000 |
| 2001 | 2001 |
| 2002 | 2002 |
| 2003 | 2003 |
| 2004 | 2004 |
| 2005 | 2005 |
| 2006 | 2006 |
| 2007 | 2007 |
| 2008 | 2008 |
Sheet1
| Year | World Wide PC Shipments | Time | World Wide PC Shipments | Fitted World Wide PC Shipments | ||||
| 81 | 2.75 | 1981 | 2.75 | 3.384 | ||||
| 82 | 5.49 | 1982 | 5.49 | 4.1076463616 | ||||
| 83 | 6.53 | 1983 | 6.53 | 4.9836726758 | ||||
| 84 | 12.5 | 1984 | 12.5 | 6.0430441788 | ||||
| 85 | 10.91 | 1985 | 10.91 | 7.3224876365 | ||||
| 86 | 13.73 | 1986 | 13.73 | 8.865309973 | ||||
| 87 | 16.47 | 1987 | 16.47 | 10.722206607 | ||||
| 88 | 19.22 | 1988 | 19.22 | 12.9519810364 | ||||
| 89 | 21.96 | 1989 | 21.96 | 15.6220536671 | ||||
| 90 | 24.705 | 1990 | 24.705 | 18.8085791372 | ||||
| 91 | 27.45 | 1991 | 27.45 | 22.5959157478 | ||||
| 92 | 30.195 | 1992 | 30.195 | 27.0750992052 | ||||
| 93 | 46.47 | 1993 | 46.47 | 32.3408724454 | ||||
| 94 | 51.96 | 1994 | 51.96 | 38.4867305399 | ||||
| 95 | 57.45 | 1995 | 57.45 | 45.5973867868 | ||||
| 96 | 65.49 | 1996 | 65.49 | 53.7381064892 | ||||
| 97 | 87.45 | 1997 | 87.45 | 62.9405660702 | ||||
| 98 | 95.49 | 1998 | 95.49 | 73.1853723349 | ||||
| 99 | 100.98 | 1999 | 100.98 | 84.3822087031 | ||||
| 100 | 134.7 | 2000 | 134.7 | 96.3497923986 | ||||
| 101 | 128.9191820837 | 2001 | 128.9191820837 | 108.7993185126 | Raw R-square (1-Residual/Total) = 0.996 | |||
| 102 | 132.4 | 2002 | 132.4 | 121.3264896586 | Mean corrected R-square (1-Residual/Corrected) = 0.990 | |||
| 103 | 169 | 2003 | 169 | 133.4179599354 | R(observed vs predicted) square = 0.990 | |||
| 104 | 189 | 2004 | 189 | 144.4772493566 | Bass Model Estimates | |||
| 105 | 207.8 | 2005 | 207.8 | 153.8722322115 | Wald Confidence Interval | |||
| 106 | 229.7 | 2006 | 229.7 | 161.0011415572 | Parameter Estimate A.S.E. Param/ASE Lower < 95%> Upper | |||
| 107 | 257.4 | 2007 | 257.4 | 165.3677583496 | M 3384.069 1303.840 2.595 620.052 6148.086 | |||
| 108 | 287.8 | 2008 | 287.8 | 166.6512970306 | P 0.001 0.000 3.795 0.001 0.002 | |||
| 109 | 28 | 164.7550347272 | Q 0.195 0.017 11.353 0.158 0.231 | |||||
| 110 | 29 | 159.8214777875 | ||||||
| 697.2 |
