Net has few degrees of separation

Researchers from Bar-Ilan University in Israel have found that the average number of connections needed to get from one point to another in real-world networks like the Internet and social networks is smaller than the number needed for randomly-connected networks.

Accounting for the difference could improve traffic flow and even provide better virus protection.

The finding has to do with the small world concept, which says that any two people in the United States are connected by less than six degrees of separation.

The researchers found that in naturally-formed networks -- like groups of people or the Internet -- the degrees of separation are fewer than in a randomly-connected network model, and this number increases extremely slowly as a network grows.

Randomly-connected network simulations are often used in designing Internet tools. The researchers' work can be used to design tools that route traffic more efficiently, improve searches, and better immunize networks against viruses. It could also be used to design networks that have shorter paths between points.

The method can be used today to develop algorithms to improve the workings of networks, according to the researchers. The work appeared in the February 7, 2003 issue of Physical Review Letters. -TECHNOLOGY RESEARCH NEWS