Scale-Free Networks"
Scientific American (05/03) Vol. 288, No. 5, P. 60; Barabasi,
Albert-Laszlo; Bonabeau, Eric
A wide array of complex systems--cellular metabolism, social networks,
and the World Wide Web, to name a few--have a shared network
architecture featuring a small number of nodes connected to a vast
number of other nodes; some nodes can support a virtually limitless
number of links, which is what makes them "scale-free." Scale-free
networks also follow predictable patterns of behavior: They can be
crippled by coordinated attacks, yet can recover easily from
accidental node failures. Research has demonstrated that scale-free
structures occur in the Internet's virtual and physical framework,
among certain sexual relationships and scientific research
collaborations, in business situations, and throughout biological
systems such as cells' protein-interaction networks. Understanding the
organizing principles of scale-free networks could be applied to many
areas, including the development of better drugs, more effective
defenses for the Internet, and controlling outbreaks of disease. Two
factors lead to the creation of scale-free hubs: Growth that gives
older nodes an advantage in acquiring links, and "preferential
attachment," whereby connections are established with a bias toward
more connected nodes. The robustness of scale-free networks is
attributed to their topological inhomogeneity--small nodes, because
they are more numerous than hubs, are more prone to failure, and the
severing of the relatively small numbers of links they support has
little effect on the network topology. However, the systems can be
disrupted if their hubs are targeted, so the best apparent solution is
to shield these hubs. Identifying Internet hubs to guard against
cyberattacks is one thing, but controlling disease epidemics is a
harder proposition, though one proposed strategy--immunizing the most
connected individuals in a population--has potential.
This archive was generated by hypermail 2b30 : Tue Jun 09 2009 - 05:00:07 EDT