,QIRUPDWLRQ ,QIUDVWUXFWXUHV Margunn Aanestad (IN364, 8. april 2002)

advertisement
Margunn Aanestad
(IN364, 8. april 2002)
,QIRUPDWLRQ
,QIUDVWUXFWXUHV
’
– Installed base
– The role of standards
– ”Network effects” (main emphasis today)
What distinguishes an II from an IS?
’ What are the implications for design
from this perspective?
7KLVOHFWXUH
’
– Star and Ruhleder (1996)
– Hanseth & Monteiro
Gore/Clinton: The Information
Superhighway – the National
Information Infrastructure Initiative
(NII), 1992.
’ IS research:
,QIRUPDWLRQ
,QIUDVWUXFWXUH
’
A perspective on large and complex
information systems.
’ Not always a sharp divide – an
information systems may have more or
less ”infrastructural characteristics”.
’ Fundamental differences between
these two approaches.
,QIRUPDWLRQ 6\VWHPYV
,QIRUPDWLRQ ,QIUDVWUXFWXUH
’
’
’
’
’
Shared resource – beyond single site use
and individual users.
Enabling - supports general use, not only
one or a few specific use areas.
Heterogeneous - a socio-technical network.
Open, no a priori definable limits
Builds on and relates to the installed base
(what is already there)
:KDWLVDQ,,"
’
Open, non-bounded Î less control
’ Sociotechnical Î complex design
’ Installed base Î less freedom in
design
’ General, shared resource Î the need
for standards
’ Network Î network effects
,PSOLFDWLRQV
’
– The installed base
– The duality of standards
– ”Network effects”
Well known ”failure stories”
’ Lack of appreciation of the role of:
'HVLJQLQJRUFKDQJLQJ
,,·VDUHFKDOOHQJLQJ
’
– The QWERTY-keyboard
– The ”Worm Community System” (S&R)
– The OSI versus IP controversy
What is already existing will influence
the new (technical and non-technical)
’ Some well-known examples:
,QVWDOOHG EDVH
’
’
’
’
To expand from a collection of bilateral
connections into a true network,
standardised interfaces are crucial.
Standards are simultaneously enabling and
constraining, it may be difficult to change
standards when it is needed.
Defining standards is a political process.
Unattainable ideals of universality,
consistency
7KHUROHRIVWDQGDUGV
1HWZRUN
Some concepts of relevance for IIs:
’ Externalities (positive and negative)
’ Increasing returns (positive feedback)
’ Standards (necessary & problematic)
’ Path dependence
’ Lock-ins
(Shapiro and Varian, 1999)
1HWZRUNHFRQRP\
•
•
Increasing returns and positive
feedback.
• Particularly relevant for network
technologies (IIs). The number of
communication partners increases the
value of the network.
E.g. when factors external to the
product itself influences its value
1HWZRUNH[WHUQDOLWLHV
Further adoptions
Reinforces value to users
Greater credibility of standard
More complements produced
Larger installed base
+RZH[WHUQDOLWLHVZRUN
’
’
’
’
Network externalities lead to path
dependency (the selfreinforcing effects
makes the network more attractive as it
grows).
Past events will have large effects on future
development.
”Early mover” advantage can influence the
whole process. (E.g. the QWERTY
keyboard, VHS vs. BETA, DVD standards).
Possible inefficiency (not the best solution)
3DWKGHSHQGHQFH
’
Lock-in: the costs of switching to a
different product/standard are too high.
’ Restricted to a not neccessarily
optimal standard (product).
’ Difficult to start a competing or
alternative network from scratch
(different standards, products)
5HVXOWORFNLQ
’
US railroad gauge is 4 feet 8.5 inches
’ English gauge for wagons – designed
for rutted roads – built by Romans to
suit horse carts for two horses.
’ The connection between a horse’s ass
and space shuttles.
([DPSOHUDLOURDGV
’
Awareness of complexity, due to the
role of the installed base and the
presence of network effects.
’ Acknowledge the emergent character,
drifting, risk, unintended effects.
Abandon control-oriented approaches?
’ What are suitable strategies for
development/change?
,PSOLFDWLRQVIRU
V\VWHPVGHYHORSHUV
’
– Radical and revolutionary change (flag days,
swift roll-outs). These conventional strategies
are possibly very problematic.
– Evolutionary strategies, e.g. ”cultivation”, are
often the only/best choice, but may be too
conservative and/or slow.
– ”Counter-networks”, build parallel alternative
networks (subnetworks) and connect through
gateways (translators).
Different strategies:
,PSOLFDWLRQVIRU
V\VWHPVGHYHORSHUV
’
’
’
Aim:Replace IPv4 with a new version for
NGI (Next Generation Internet).
IPv6 was selected among several
candidates. An important criteria for the
choice was the WUDQVLWLRQVWUDWHJ\. IPv6
made a gradual transition possible (dual
stack etc.)
– Limited address space (8bits)
– No support for mobility, security, QoS etc.
Limitations of IP version 4 (current protocol):
([DPSOH7&3,3
’
’
Wednesday April 10th: K.R. on design and
implementation of global information
systems using traditional SD methods.
Focus on the problem of standards and the
installed base.
Monday April 15th: M.A. on telemedicine:
Evolutionary development and its
challenges. ”Bootstrapping” the
development process.
WREHFRQWLQXHG
'HVLJQDQGGHYHORSPHQWRI
,QIRUPDWLRQ ,QIUDVWUXFWXUHV
’
’
’
Hanseth, Monteiro and Hatling (1996): Information
Infrastructure Development: The Tension between
Standardization and Flexibility. 6FLHQFHWHFKQRORJ\
DQG+XPDQ9DOXHVvol. 21, no. 4, pp. 407-426.
Shapiro and Varian (1999): ,QIRUPDWLRQ5XOHV± D
VWUDWHJLFJXLGHWRWKHQHWZRUNHFRQRP\ Harvard
Business School Press.
Star and Ruhleder (1996): The Ecology of
Infrastructures: Problems in the Implementation of
Large-scale Information Systems. ,QIRUPDWLRQ
6\VWHPV5HVHDUFK, vol. 7., no.1, pp. 111-134.
5HIHUHQFHV
Download