A Rationale for International Cooperation Brian Randell

1
A Rationale for International Co-operation?
Brian Randell
2
Two Laws that shape the Future

• Moore The number of transistors per chip will
  double every eighteen months - this law has
  enabled the growth of the internet
• Metcalfe "The usefulness of a network varies as
  the square of the number of users - this law has
  made the growth actually happen

3
Three Laws that really shape the Future

• Moore The number of transistors per chip will
  double every eighteen months
• Metcalfe The usefulness of a network varies as
  the square of the number of users.
• Murphy If anything can go wrong, it will.

4
The Primacy of Socio-Technology

• each of the 3M laws is socio-technological, not
  merely technological
• most of the major systems we need to concern
  ourselves with are socio-technological, i.e.
  computer-based systems involving people as well
  as computers and networks
• many of their problems have socio-technological
  causes, whose solutions need to be grounded in
  (good) socio-technology.

5
The Relevance of International Co-operation

• we face a future characterized by dependence on
  global (socio-technological) systems, and
  systems-of-systems
• the design and operation of such S-T systems
  needs to allow for the social heterogeneity
  across nations (among system users and mis-users)
• such problems, being international in both their
  scope and their variety, demonstrate the need for
  international co-operation on system
  dependability research, if they are to be
  satisfactorily solved

6
Social Heterogeneity

• attitudes to risk vary and evolve (will
  computer-based systems crashes be treated as
  casually as automobile crashes or as seriously as
  plane crashes?)
• government policies differ e.g. relating to
  balances between personal privacy and national
  security, and between (multi-national) companies
  and the state
• there are horrendous economic imbalances
  globally e.g. from shared clockwork radios to
  personal electronic bazaars

7
Possible Priority Topics

• the design of intrusion-tolerant global systems
• dependability of global systems of systems
• the problems of subdividing responsibility
  between humans and computers in global
  computer-based systems
• (so far, a typical researchers list - i.e.
  essentially incremental)
• inherently-dependable systems
  (self-stabilization writ large?)
• global systems whose interfaces and
  specifications are ill-defined and/or
  continuously evolving

8
A new dependability research opportunity

• the GRID - a (very) well-funded global
  (socio-technological!) system, originated by the
  high energy physics community, initially
  advertised as a successor to the internet and the
  Web!
• first aimed at access to massive computing, like
  Arpanet was initially, now aimed at supporting
  virtual organizations
• Uses Linux plus Globus middleware
• IBM, Sun, etc., joining in (in part as a riposte
  to MSs .net?)
• in US, relatively little involvement of major CS
  departments, but in UK (and France?) they are
  becoming involved
• it raises major (short and long term)
  dependability issues

See The Anatomy of the Grid Enabling Scalable
Virtual Organizations. I. Foster, C. Kesselman,
S. Tuecke, Intl. J. Supercomputer Applications,
15(3), 2001. http//www.globus.org/research/papers
/anatomy.pdf And for a UK view
http//e-science.ox.ac.uk/events/19-sep-2001/hey.h
tm